+ All Categories
Home > Documents > MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX...

MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX...

Date post: 30-Jun-2020
Category:
Upload: others
View: 3 times
Download: 0 times
Share this document with a friend
127
MODELING CO 2 SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou Dissertation submitted to the faculty of the Virginia Polytechnic Institute and State University in partial fulfillment of the requirements for the degree of Doctor of Philosophy In Mining Engineering Michael E. Karmis, Co-Chair Nino S. Ripepi, Co-Chair Zacharias G. Agioutantis Gerald H. Luttrell Kramer D. Luxbacher May 06, 2014 Blacksburg, VA Keywords: coalbed methane, carbon dioxide, reservoir modeling, unconventional reservoirs ©2014, Foteini Vasilikou
Transcript
Page 1: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

MODELING CO2 SEQUESTRATION AND ENHANCED

GAS RECOVERY IN COMPLEX UNCONVENTIONAL

RESERVOIRS

Foteini Vasilikou

Dissertation submitted to the faculty of the Virginia Polytechnic Institute and State

University in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

In

Mining Engineering

Michael E. Karmis, Co-Chair

Nino S. Ripepi, Co-Chair

Zacharias G. Agioutantis

Gerald H. Luttrell

Kramer D. Luxbacher

May 06, 2014

Blacksburg, VA

Keywords: coalbed methane, carbon dioxide, reservoir modeling, unconventional

reservoirs

©2014, Foteini Vasilikou

Page 2: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

MODELING OF CO2 SEQUESTRATION AND ENHANCED GAS

RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS

Foteini Vasilikou

ABSTRACT

Geologic sequestration of CO2 into unmineable coal seams is proposed as a way to

mitigate the greenhouse gas effect and potentially contribute to economic prosperity

through enhanced methane recovery.

In 2009, the Virginia Center for Coal and Energy Research (VCCER) injected 907

tonnes of CO2 into one vertical coalbed methane well for one month in Russell County,

Virginia (VA). The main objective of the test was to assess storage potential of coal

seams and to investigate the potential of enhanced gas recovery. In 2014, a larger scale

test is planned where 20,000 tonnes of CO2 will be injected into three vertical coalbed

methane wells over a period of a year in Buchanan County, VA.

During primary coalbed methane production and enhanced production through CO2

injection, a series of complex physical and mechanical phenomena occur. The ability to

represent the behavior of a coalbed reservoir as accurately as possible via computer

simulations yields insight into the processes taking place and is an indispensable tool for

the decision process of future operations. More specifically, the economic viability of

projects can be assessed by predicting production: well performance can be maximized,

drilling patterns can be optimized and, most importantly, associated risks with operations

can be accounted for and possibly avoided.

However, developing representative computer models and successfully simulating

reservoir production and injection regimes is challenging. A large number of input

parameters are required, many of which are uncertain even if they are determined

experimentally or via in-situ measurements. Such parameters include, but are not limited

to, seam geometry, formation properties, production constraints, etc.

Modeling of production and injection in multi-seam formations for hydraulically

fractured wells is a recent development in coalbed methane/enhanced coalbed methane

(CBM/ECBM) reservoir modeling, where models become even more complex and

demanding. In such cases model simulation times become important.

Page 3: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

iii

The development of accurate simulation models that correctly account for the

behavior of coalbeds in primary and enhanced production is a process that requires

attention to detail, data validation, and model verification. A number of simplifying

assumptions are necessary to run these models, where the user should be able to balance

accuracy with computational time.

In this thesis, pre- and post-injection simulations for the site in Russell County, VA,

and preliminary reservoir simulations for the Buchanan County, VA, site are performed.

The concepts of multi-well, multi-seam, explicitly modeled hydraulic fractures and skin

factors are incorporated with field results to provide accurate modeling predictions.

Page 4: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

iv

DEDICATION

This dissertation is dedicated to my beloved mother Alkioni Vasilikou, my father

Constantinos Vasilikos and my brother Ioannis Vasilikos, M.D., Ph.D., who never left my side

and taught me that “where there’s a will, there’s a way.”

Page 5: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

v

ACKNOWLEDGMENTS

There are no words to express my gratitude to all the people who have supported me in this

challenging period of my life. I sincerely thank Professor Michael Karmis for his guidance and

advice in all aspects of my educational and professional life. Without his support I wouldn’t be

here.

I would like to thank Dr. Nino Ripepi for his support during the past three years. He has

been an excellent advisor and teacher and helped me overcome many difficulties along this

journey.

I cannot thank enough Professor Zach Agioutantis for his skill, teaching, motivation,

guidance and patience. Professor Agioutantis has been an amazing mentor.

I would like to express my gratitude to Dr. Kray Luxbacher for her advice and guidance, and

to Professor Gerry Luttrell who was the first person to welcome me to the Mining Engineering

Department and guide my studies.

A special thank you to Steve Schafrik, whose exceptional technical and coding skills were of

utmost importance in this effort.

I would also like to thank Dr. Cigdem Keles for her valuable help and support with the

reservoir modeling.

Finally, a big “Thank you” to everyone at the Virginia Center for Coal and Energy Research

who has contributed with their support in this work and especially my dear friend and advisor

Dr. John Craynon.

I would also like to express my love for my amazing mother Alkioni, who never left my side

and crossed the Atlantic Ocean numerous times to be by my side; to my cousin, Dr. Demetrios

Vavvas, who has been a father, an advisor and a friend to me; and to my brother Ioannis for his

constant love and support.

I am thankful to my friends Christos, Constantinos and Korina for always being there for

me.

Page 6: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

vi

TABLE OF CONTENTS

Abstract ____________________________________________________________________ ii

Dedication __________________________________________________________________ iv

Acknowledgments ____________________________________________________________ v

Table of Contents ____________________________________________________________ vi

List of Tables ________________________________________________________________ ix

List of Figures _______________________________________________________________ x

Preface ____________________________________________________________________ xiii

The Application of Constitutive Laws to Model the Dynamic Evolution of Permeability in

Coal Seams for the Case of CO2 Geologic Sequestration and Enhanced Coal Bed Methane

Recovery____________________________________________________________________ 1

Abstract _______________________________________________________________ 1

Introduction ____________________________________________________________ 2

Coalbed Characteristics ___________________________________________________ 3

Coalbed Methane and Enhanced Coalbed Methane Production Mechanisms _________ 5

Permeability Models _____________________________________________________ 6

Discussion on Implementation of Permeability Models for Reservoir Simulation _____ 17

Conclusions ___________________________________________________________ 19

Acknowledgements _____________________________________________________ 20

Nomenclature __________________________________________________________ 20

References ____________________________________________________________ 21

Experiences in Reservoir Model Calibration for Coal Bed Methane Production in Deep

Coal Seams in Russell County, Virginia _________________________________________ 27

Abstract ______________________________________________________________ 27

Introduction ___________________________________________________________ 28

Geology ______________________________________________________________ 30

Well Stimulation _______________________________________________________ 31

Page 7: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

vii

Production History ______________________________________________________ 34

Modeling _____________________________________________________________ 35

Description of Production Model ________________________________________ 35

Results and Discussion __________________________________________________ 40

Conclusions ___________________________________________________________ 44

Acknowledgments ______________________________________________________ 45

References ____________________________________________________________ 45

Model Verification of Carbon Dioxide Sequestration in Unmineable Coal Seams with

Enhanced Coalbed Methane Recovery __________________________________________ 47

Abstract ______________________________________________________________ 47

Introduction ___________________________________________________________ 48

Field Description _______________________________________________________ 49

Field Site Layout _____________________________________________________ 50

CO2 Injection __________________________________________________________ 50

Injection Logging _____________________________________________________ 51

Monitoring Well Results _______________________________________________ 52

Reservoir Modeling _____________________________________________________ 52

Theoretical Considerations for the Skin Factor and

the Hydraulic Fracture Approach ______________________________________ 53

The Importance of Relative Permeability __________________________________ 54

History Match _______________________________________________________ 55

Results and Discussion __________________________________________________ 56

Conclusions ___________________________________________________________ 59

Acknowledgments ______________________________________________________ 59

References ____________________________________________________________ 60

Reservoir Simulations for Coal Bed Methane (CH4) Production and Carbon Dioxide (CO2)

Injection in Deep Coal Seams in Buchanan County, Virginia _______________________ 62

Reservoir Model Calibration for Coal Bed Methane Production

from Deep Coal Seams in Buchanan County, Virginia ______________________ 62

Abstract ______________________________________________________________ 62

Page 8: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

viii

Introduction ___________________________________________________________ 63

Geology ______________________________________________________________ 63

Well Stimulation _______________________________________________________ 65

Production History ______________________________________________________ 65

Description of Reservoir Model ___________________________________________ 67

18- Layer Models _____________________________________________________ 68

5-Zone Model ________________________________________________________ 73

Results and Discussion __________________________________________________ 74

Initial Volumetrics ____________________________________________________ 74

Production Mechanism ________________________________________________ 79

Conclusions ___________________________________________________________ 81

CO2 Injection Model for Enhanced Coalbed Methane

Recovery in Deep Coal Seams in Buchanan County, VA _____________________ 82

Abstract ______________________________________________________________ 82

Development of Reservoir Models _________________________________________ 82

Results and Discussion __________________________________________________ 89

S Models ____________________________________________________________ 90

L Models ___________________________________________________________ 96

Conclusions __________________________________________________________ 103

Acknowledgements ____________________________________________________ 103

References ___________________________________________________________ 104

Summary and Conclusions __________________________________________________ 107

Further Work _________________________________________________________ 110

Recommendations for Transferring Lessons Learned

from Coalbed Methane Modeling to Shale Gas Modeling ____________________ 111

Page 9: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

ix

LIST OF TABLES

Table 1 - BD114 Stimulated Coal Seams (VaDMME, 2002) __________________________ 33

Table 2 - Parameters and Material Properties Used in the Reservoir Model _______________ 37

Table 3 - Optimal Parameters Determined for the Dingle-Well Runs ____________________ 40

Table 4 - Parameter Range Used as Input in the CMOST Model Runs ___________________ 42

Table 5 - Optimal Parameters After CMOST Runs and Manual Adjustments for the Multi-Well

Runs ______________________________________________________________________ 43

Table 6 - Parameters and Material Properties Used in the 18-Layer Reservoir Model _______ 70

Table 7 - Set of Parameters Varied in Simulations for Cases C1 to C11 __________________ 71

Table 8 - Input Parameters for Modeling Hydraulic Fractures in Scenarios H1, H2, H3, H4 __ 73

Table 9 - Parameters and Material Properties Used in the 5-Zone Reservoir Model _________ 74

Table 10 - Reference Names and Production Start Dates for the Coalbed Methane Wells Included

in the Simulations ____________________________________________________________ 83

Table 11 - Layers Number, Coal Seam Names, Thicknesses and Top Elevations for the Geologic

Structure Assumed for the Study Area ____________________________________________ 85

Table 12 - Input Parameters for the Combined and Four-Well Models ___________________ 86

Table 13 - CO2 Flowback at the Injectors and CO2 Breakthrough at the Offset Well for the (S)

Models for Projection Time to Year 2023 _________________________________________ 95

Table 14 - Cumulative CH4 Production for the (S) Models for the Base (S1), Skin (S2) and

Hydraulic Fractures (S3) Scenarios for Projection Time to Year 2023 ___________________ 95

Table 15 - Cumulative Gas Production (CGP) for All Wells for Cases, L1a, L1b, L2a, L1c _ 101

Table 16 - CO2 Flowback and Breakthrough for Cases, L1a, L1b, L2a, L1c ______________ 102

Page 10: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

x

LIST OF FIGURES

Figure 1 - Face and Butt Cleats___________________________________________________ 4

Figure 2 - The Matchstick Model _________________________________________________ 8

Figure 3 - Different Modes of Coupled Simulation (a, b, c) and Uncoupled Simulation (d) ___ 10

Figure 4 - Uniaxial Strain Model ________________________________________________ 11

Figure 5 - Fracture Aperture Changes with Swelling/Shrinkage ________________________ 11

Figure 6 - Matchstick Model with Bridging Between Coal Blocks ______________________ 16

Figure 7 - Domain Discretization ________________________________________________ 18

Figure 8 - Case Study Area Showing Injection Wells ________________________________ 18

Figure 9 - Initial Attempts for Coupled Simulation Using an External Geomechanics Package

(Vasilikou et al., 2012) ________________________________________________________ 19

Figure 10 - Location of Field Test Site (Ripepi et al, 2009) ____________________________ 29

Figure 11 - Layout of Injection and Offset Wells (Karmis et al., 2008) ___________________ 29

Figure 12 - (A) Rose Diagram Showing Face and Butt Cleat Orientation (B) Rotation by 35o

Clockwise __________________________________________________________________ 31

Figure 13 - Gas Production History at the Injection and the Offset Wells _________________ 34

Figure 14 - 3D View of Model Grid and Wells _____________________________________ 36

Figure 15 - Relative Permeability Curves Used in the Single Well Runs (Gash et al. 1993, Mavor

et al. 1953) (SWT: Water Relative Permeability or Water Saturation Fraction; SLT: Gas Relative

Permeability or Gas Saturation Fraction) (see Table 2) _______________________________ 38

Figure 16 - Face Cleat Permeability Regions for Single-Well Models and Combined Model _ 39

Figure 17 - Butt Cleat Permeability Regions for Single-Well Models and Combined Model __ 39

Figure 18 - History Match on Gas and Water Rates for Well BD-115 (RU-132) Under a Single

Well Run. The Bottom Hole Pressure is Also Shown ________________________________ 41

Figure 19 - History Match of Gas and Water Rates for Well BD-115 (RU-132) for the Combined

Run. The Bottom Hole Pressure is Also Shown _____________________________________ 41

Figure 20 - Relative Permeability Curves as Determined by the Optimization Algorithm for Two

of the Eight Production Wells (SWT: Water Relative Permeability or Water Saturation Fraction;

SLT: Gas Relative Permeability or Gas Saturation Fraction) ___________________________ 43

Page 11: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

xi

Figure 21 - History Match on Gas and Water Rates for Well BD-115 (RU-132) for the Combined

Run with Optimized Input Parameters. The Bottom Hole Pressure is Also Shown __________ 44

Figure 22 - (A) Rose Diagram Showing Face and Butt Cleat Orientation (B) Rotation by 35o

degrees clockwise ____________________________________________________________ 51

Figure 23 - Typical Relative Permeability Curves for Two Wells _______________________ 54

Figure 24 - Relative Permeability Curves for the HF and SF Models ____________________ 55

Figure 25 - Face Cleat Permeability Variation for the Base Model ______________________ 56

Figure 26 - Typical HF Discretization for All Wells _________________________________ 56

Figure 27 - Cumulative Water Production vs. Measured Data for Well BC-115 (RU-123) for All

Models_____________________________________________________________________ 57

Figure 28 - Gas Rate and Cumulative Gas during CO2 Injection into Well BD 114 (RU 84) __ 58

Figure 29 - CO2 Plume around Injection Well Immediately After Injection Has Been Completed

In the HF Model _____________________________________________________________ 58

Figure 30 - Daily Gas Production Rate of the Injection Wells __________________________ 66

Figure 31 - Daily Water Production Rate of the Injection Wells ________________________ 67

Figure 32 - (A) 18-Layer Model, (B) 5-Zone Model _________________________________ 68

Figure 33 - History Matching of Monthly Gas Rate and Cumulative Gas for Well DD7 Up to

Year 2013 __________________________________________________________________ 69

Figure 34 - Langmuir Adsorption Isotherms for Langmuir Volume Constants 500, 650, 800

scf/ton and Langmuir Pressure Constants 100 and 333 psi; Maximum and Minimum Pore

Pressures in the Reservoir for Pressure Gradients 0.315 and 0.36 psi/ft __________________ 71

Figure 35 - Different Scenarios for the Relative Permeability Curves to Water and Gas _____ 72

Figure 36 - Original Gas in Place in the Model and Initially Adsorbed Gas on the Coal Matrix for

All Cases (C) ________________________________________________________________ 76

Figure 37 - Original Water in Place for All Cases (C) ________________________________ 76

Figure 38 - Coal Matrix Volume per Layer Starting from the Shallowest (Layer 1) to the Deepest

(Layer 18) Seam for Cases C3 and C11 ___________________________________________ 77

Figure 39 - Minimum and Maximum Pore Pressures per Layer for Two Different Pressure

Gradients Used In C Cases _____________________________________________________ 77

Figure 40 - Initially Adsorbed Gas per Layer for Cases C3, C5 and C8 __________________ 78

Figure 41 - Originally Adsorbed Gas per Layer for the 5-Zone Model ___________________ 80

Page 12: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

xii

Figure 42 - Originally Adsorbed Gas per Layer for the 18-Layer Model _________________ 81

Figure 43 - Relative Permeability Curves to Water and Gas as Determined Through History

Matching Gas Production for All the Wells in the Study Area _________________________ 89

Figure 44 - Gas Rate at Surface Conditions for the Three Injectors, DD7, DD7A, and DD8

throughout History Matching, Injection and Forecasting to Year 2023 ___________________ 90

Figure 45 - (A) Cumulative Gas Production per Layer Up to Year 2013 and (B) CO2 Adsorption

per Layer during Injection and Post Injection Up to Year 2023 for Scenario S3 ____________ 93

Figure 46 - CO2 Adsorption Profile in gmole/ft3

at Layer 4 for Scenarios (A) S1 Base, (B) S2

Skin and (C) S3 Hydraulic Fractures Scenario ______________________________________ 94

Figure 47 - Gas Mass Rate of Injected CO2 at Surface Conditions Versus Time for the Injection

Scenarios L1 and L2 __________________________________________________________ 97

Figure 48 - Pressure-Temperature Phase Diagram for CO2 and Bottom Hole Pressures at

Injection Well DD7 for Injection Plans L1a, L1b, and L1 _____________________________ 98

Figure 49 - Cumulative Gas Production of All the Wells in the Study Area for All Cases

Examined for the (L) Models ___________________________________________________ 99

Page 13: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

xiii

PREFACE

This research effort comprises the four major tasks summarized below. These tasks are

addressed in a set of scholarly works which are either published or will be submitted for

publication.

Task 1. Investigation of dynamic evolution permeability models

The main objectives for Task 1 are to:

a) Critically review permeability change models which are proposed in the literature for

coalbeds with respect to primary and enhanced production.

b) Implement a permeability change model in single well models by coupling a

reservoir simulator and appropriate geomechanical code.

Task 2. Sensitivity analysis for numerical models

The main objectives for Task 2 are to:

a) Perform sensitivity analyses to study the behavior of the reservoir models when

production and injection occur in multiple zones.

b) Examine the response of the reservoir models with respect to varying input

parameters that affect the volumetrics properties.

c) Validate well and reservoir characteristics in the models that have an effect on the

production mechanism.

Task 3. Comparison of different approaches that model well stimulation in reservoir

simulations

The main objectives for Task 3 are to:

a) Develop models where a negative skin factor is assigned.

b) Explicitly model hydraulic fractures in reservoir models.

c) Study and compare enhanced flow properties and the effect on injection depending

on the well stimulation approach.

Task 4. Predict enhanced methane recovery, CO2 flowback and breakthrough

The main objectives for Task 4 are to:

Page 14: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

xiv

a) Develop full field scale models and incorporate all key modeling elements for

coalbeds.

b) History-match the final models and create the final baseline model.

c) Predict enhanced methane recovery at the injectors and offset wells.

d) Predict CO2 plumes and potential breakthrough.

Page 15: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

1

THE APPLICATION OF CONSTITUTIVE LAWS TO MODEL THE

DYNAMIC EVOLUTION OF PERMEABILITY IN COAL SEAMS FOR

THE CASE OF CO2 GEOLOGIC SEQUESTRATION AND ENHANCED

COAL BED METHANE RECOVERY1

Foteini Vasilikou, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Nino Ripepi, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Zach Agioutantis, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Michael Karmis, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Abstract

Injection and storage of carbon dioxide (CO2) in deep unmineable coalbeds decreases

anthropogenic greenhouse gas emissions and presents a financially viable solution by enhancing

recovery of coalbed methane (ECBM). Coalbeds are commonly characterized by a dual porosity

system, which is comprised of a network of natural fractures (cleats) and matrix blocks of coal

exhibiting highly heterogeneous porosity. The gas transport through the cleat system is governed

by Darcy's Law. This study reviews and critically evaluates available models for describing

coalbed permeability that can be applied to calculate gas flow in such systems. In addition, the

1 The Application of Constitutive Laws to Model the Dynamic Evolution of Permeability in

Coal Seams for the Case of CO2 Geologic Sequestration and Enhanced Coal Bed Methane

Recovery. F. Vasilikou, N. Ripepi, Z. Agioutantis, M. Karmis.

This paper has been published already in the proceeding of the 29th Pittsburgh Annual Coal

Conference in 2012.

Foteini Vasilikou researched and prepared this manuscript, with Nino Ripepi, Zach

Agioutantis and Michael Karmis providing technical and editorial input.

Page 16: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

2

potential of using geomechanical models to better account for the physical processes that occur

during coalbed methane production and CO2 injection and storage is also investigated. The

results of this review can be used for evaluating modeling approaches when employing reservoir

simulators to simulate injection and storage in ECBM cases.

Introduction

Historically, recovery of natural gas or coalbed methane (CBM) is described as early as

1858. By 1953 degasification of the Pittsburgh seam was underway for safety reasons. By 1980,

companies began profiting from deep coal seams where natural gas was of interest as a clean and

domestic energy source, whether the coal was mineable or not (Bodden and Ehrlich, 1998;

Ayers, 2002; Zarrouk, 2008; Liu, 2011). Coal seams are termed as unmineable where mining is

considered to be infeasible given foreseeable technology, costs, sales prices, inadequate seam

thickness, poor areal continuity, adverse geology (steeply dipping, rolls, faults), poor coal

quality, excessive depths and other reasons. Typical depths for unmineable coal seams are 300-

900 m (Siriwardane, 2008). The most methane that could be recovered by the pressure depletion

method is not anticipated to be larger than 50% of gas-in-place, even after several decades of

production (Puri and Yee, 1990). Hence, in the 90’s enhanced coalbed methane (ECBM)

production via injection of CO2 in the unmineable coal seams was suggested as a more efficient

way of extracting a larger fraction of methane in place without having to overly reduce reservoir

pressure. In 2005 under the Kyoto Protocol, an international agreement linked to the United

Nations framework convention on climate change, a cap on reduction of greenhouse gas

emissions 5% below the 1990 greenhouse levels, further increased the interest in sequestering

CO2 in coalbeds. Thus, enhanced coalbed methane has the twofold benefit of enhanced

production and redeeming greenhouse gases pollution. In Australia, Canada, China, Poland and

the USA, a number of ECBM projects are ongoing (Hamelinck et al., 2003; Law et al., 2002;

Damen et al., 2005; Zarrouk, 2008).

Large ‘natural gas’ or methane reserves are existent all over the world such as in Canada,

Russia, China, the United States and Australia.

There exist a number of physical processes which occur during production of coalbed

methane and injection and storage of CO2 in coal seams. Coalbeds are characterized by a dual

porosity system, which is comprised of a network of natural fractures (cleats) and matrix blocks

Page 17: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

3

of coal exhibiting highly heterogeneous porosity. The gas transport through the coal matrix

micropores (primary porosity), driven by the concentration gradient, is controlled by Fickian

diffusion. The flow of gas and water through the cleat system (secondary porosity) is governed

by Darcy's Law. Methane is initially stored in the adsorbed state, within the porous structure of

the coal matrix blocks, which is usually described by a Langmuir-type isotherm. During coalbed

methane production, reservoir pressure is decreased through dewatering, allowing methane

molecules to desorb from the internal coal matrix (Fickian diffusion) and travel through the cleat

structure (Darcy Flow). The opposite process occurs during CO2 injection. This results in an

inverse diffusion process, where CH4 molecules are desorbed from the coal matrix and replaced

by CO2 molecules. The combination of reservoir dewatering and its associated depressurization,

CH4 desorption and CO2 adsorption causes alterations in the stress regime acting on the coal

matrix. Gains or losses in water and gas-relative permeability can be noted, but vary greatly in

accordance with the geological characteristics of the coal.

Coalbed Characteristics

Coal is a highly heterogeneous porous medium that contains micropores (<2 nm), mesopores

(between 2 and 50 nm), macropores (>50 nm) and natural fractures formed during coalification.

(Wang 2009; Shi and Durucan, 2005). In the literature, coalbeds or coal seams are characterized

by a dual-continuum system comprised by the porous coal matrix and cleats (fractures) (Liu and

Rutqvist, 2009). Cleats are considered to be a system of densely spaced, orthogonal fractures;

cleats are comprised of face and butt cleats. Face cleats are long, well-developed, almost planar

fractures that extend parallel to each other. Butt cleats, run at an angle of roughly 90 degrees to

the face cleats and usually terminate at them (Figure 1). Face and butt cleats are approximately

perpendicular to the bedding planes (Liu 2011; Gu 2009).

Page 18: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

4

Figure 1 - Face and Butt Cleats

Face and butt cleats in conjunction with larger scale fractures, joints and faults constitute

effective flow paths for gas in the coalbeds. As it has been postulated and confirmed by many

experimental results, the anisotropic pore structure of coal results in higher permeability values

along the face cleats compared to butt cleats and significantly higher than in the direction normal

to the bedding plane. Indicative permeability values reported by Gash et al. (1992) are 0.6~1.7

mD for face cleats, 0.3~1 mD for butt cleats and only 0.007 mD for the direction normal to the

bedding plane (Gu, 2009). The porous coal matrix permeability, is reported to be substantially

less than that for the cleats system, normally eight times smaller; thus many researchers

disregard coal matrix permeability and accredit the coalbed permeability to its cleats and larger

scale discontinuities (Robertson, 2005; Liu and Rutqvist, 2009).

Porous coal matrix is the primary medium for gas storage in coalbeds and accounts for up to

95-98% storage, (Gray, 1987). Storage in the coal matrix is performed via two primary

mechanisms. More specifically, gas can either be physically adsorbed on the very large (20-200

m2/g) (Patching, 1970) internal surface of the porous coal matrix, or it can be adsorbed within

the molecular structure of the coal matrix. The remaining gas resides as either a free gas or

dissolved in water within the cleats and larger fractures (Marsh, 1987; Shi and Durucan, 2005).

Typically a Langmuir-type adsorption isotherm is employed to describe the adsorption

phenomenon in coals (Shi and Durucan 2005).

In accordance with their basic characteristics, coalbeds can be regarded as dual porosity and

single permeability systems (Harpalani and Schraufnagel, 1990; Lu and Connell, 2007; Warren

and Root, 1963; Liu, 2011).

With natural gas production (CBM) and CO2 storage in unmineable coalbeds (ECBM),

complex interactions with varying stress state and sorption phenomena are triggered that affect

Page 19: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

5

the transport and sorptive properties of coal and consequently affect production and/or injectivity

rates. Therefore, understanding the mechanisms of the dynamic evolution of sorption, flow, coal

deformation, porosity and permeability is of fundamental importance to CBM/ECBM recovery

(Liu 2011).

Coalbed Methane and Enhanced Coalbed Methane Production Mechanisms

Extraction of methane from undersaturated coalbeds activates a sequence of interactions

among the porous coal matrix, fractures and adsorbed and free gas. In primary gas production,

methane is extracted from the coalbed via decreasing reservoir pressure. More specifically, by

dewatering the targeted seams, the pressure of the gas in the reservoir decreases which results in

an increase of the effective stress and consequently closure of fracture apertures and lower

permeability. Furthermore, when the gas pressure falls below the desorption point, methane is

liberated from the matrix into the cleats system and the matrix shrinks. Based on the zero volume

change condition, matrix shrinkage effectively causes the aperture of fractures to increase and

thus fracture permeability to increase. Hence, an initial decrease in fracture permeability

attributed to an increase in effective stress is counteracted by a gradual increase in permeability

attributed to matrix shrinkage. The net permeability depends on the net impact of these two

competing effects (Chen et al., 2008; Connell, 2009; Liu et al., 2010b, c, d; Shi and Durucan,

2004; Gu, 2009; Liu, 2011).

Coalbed methane extraction for saturated coal seams follows the same mechanisms as in the

case for undersaturated coals without the dewatering phase. When pressure starts dropping there

is immediate release of methane from the coal matrix to the fracture system. Finally for

oversaturated coal seams, reduction of reservoir pressures results in concurrent flow of mobile

water and free gas towards the production wells. As soon as the coalbed seam attains a state of

saturation the production mechanism is similar to that for saturated coal seams (Gu 2009).

Carbon dioxide has been shown to have greater affinity to the porous coal matrix in contrast

to methane and thus it is preferentially adsorbed onto the coal when present. According to

laboratory measurements it is indicated that depending on coal rank, the matrix can adsorb

almost twice the amount of carbon dioxide by volume as methane and in some cases of lower

coal rank this ratio may even be as high as 10 to 1 under conditions of normal reservoir pressure

(Stanton et al., 2001; Shi and Durucan, 2005). In ECBM, when carbon dioxide is competitively

Page 20: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

6

adsorbed onto the coal matrix and methane is displaced the coal matrix swells. Coal matrix

swelling combined with the zero volume change condition results in reduced fracture apertures

and decreased fracture permeability (Mavor et al, 2002; Gu, 2009).

Permeability Models

As discussed, permeability is a key parameter in both coalbed and enhanced coalbed

methane production. US experience suggests that an absolute permeability of 1 mD is generally

required to achieve commercial production rates. For these reasons, the complexity of the

physical interactions affecting permeability has been extensively discussed in the literature and a

variety of models have been proposed attempting to simulate the dynamic evolution of this

phenomenon. Many of these models have also been incorporated in numerical simulators in

order to quantify coal-gas interactions (Liu 2011). All of these models are based on numerous

assumptions regarding flow characteristics and boundary conditions as well as the

geomechanical component that may interact with the above parameters. In the following section

the most important models representing the key parameters in multiphase flow problems are

presented and critically evaluated. Also, in order to achieve a better understanding of the

different mechanisms involved as well as the interaction of the various parameters of the model,

formal definitions of the important variables are briefly outlined below.

A porous medium having a bulk volume (V), comprised of solid volume (Vs) and pore

volume (Vp), where V=Vs+Vp, has porosity (φ) which is defined as the ratio of the volume of

the pores to the bulk volume of the medium, φ=Vp/V. The permeability (k) of the porous

medium is the measure of its ability to allow fluids to pass through it and is expressed in units of

area. Compressibility of a solid medium is described as the measure of relative volume change as

a response to a pressure or a mean stress change. Furthermore, volumetric strain of a deformed

body is defined as the ratio of the change in volume of the body to the deformation of its original

volume. Effective stresses (and or strains) can be calculated from total quantities by subtracting

the fraction attributed to pore pressure.

Physical processes are complex phenomena and their formal descriptions, which are also

known as constitutive laws, are usually very complicated. Incases simplifying assumptions are

used to facilitate analyses with a first order approximation. In the following, it will be

demonstrated that researchers may apply different constitutive equations to describe the same

Page 21: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

7

physical process with a varying number of (simplifying) assumptions. The basic principle of

each formulation will be presented along with the underlying assumptions. On many occasions

the governing equations of these physical processes are termed analytical models in contrast to

the numerical models, which include the methodology for achieving a numerical solution to the

analytical equation.

Liu 2011 stipulated that the total effective volumetric strain responsible for changes in coal

porosity and permeability during production of CBM/ECBM as described by a thermo-poro-

elastic constitutive equation is given by:

(1)

It should be noted that the basic assumptions for this formulation are (i) that thermal

contraction of a non-isothermal body is analogous to coalbed matrix shrinkage/swelling (Palmer

and Mansoori, 1996), and (ii) that the coal seam is treated as a non-isothermal medium.

According to Eq. (1) the total effective volumetric strain is comprised of the total volumetric

strain, the coal compressive strain, the gas sorption-induced volumetric strain and the thermal

strain.

This model is considered to include all major terms affecting fluid flow in single and dual

porosity situations affected by the strain regime of the medium.

Porosity and permeability of the coalbed can be both described as a function of the total

effective volumetric strain ( ) through the following general functions:

( ) (2)

( ) (3)

Also, formulations relating porosity and permeability have been reported in the literature for

different conceptual permeability models for coalbeds. Chilingar (1964) regarded coal as a

continuum porous medium and proposed an equation relating permeability to the porosity and

the effective diameter of its grains, Eq. (4).

( ) (4)

Page 22: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

8

By applying Eq. (4) for initial ( ) and current ( , k) permeability and porosity values,

Eq. (5) can be derived. When the porosity is much smaller than 1 the second term of the right-

hand side is close to unity and the term is dropped. Hence, the relationship between permeability

and porosity for the coal matrix is as follows:

(

)

(

)

(

)

(5)

Reiss (1980) idealized coal as a collection of matchsticks (Figure 2) and considered

permeability to be related to its porosity and cleat system spacing, Eq. (6).

(6)

By applying Eq. (6) for initial ( ) and current (k) permeability values, Eq. (7) can be

derived.

(

)

(7)

Figure 2 - The Matchstick Model

In Eq. (7) which is the same as the simplified version of Eq. (5), the value of the exponent is

related to the flow regime, cleat size and roughness and is normally equal to 2 for laminar flow, 3

for transition flow and 4 for turbulence flow. According to the literature, flow through naturally

fractured coalbed reservoirs is in the transition flow regime, and thus the exponent is taken to be

3 (Reiss, 1980; Seidle, 1992; Avraam and Payatakes, 1995; Wang, 2009).

By applying Eq. (7) which relates permeability to porosity (and thus to elemental volume

conditions), permeability-change formulations are acquired which in many cases may include

coal properties as well as related physical quantities. Extended research has been conducted and

Coal

Confining formation

Page 23: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

9

many models have been proposed in the past 30 years in order to combine the flow regime to the

geomechanical characteristics of the medium and examine the dynamic evolution of

permeability. Different classifications have been proposed for these models according to their

respective assumptions, encompassed parameters and employed solving techniques.

More specifically, Palmer (2007) grouped permeability models into analytical permeability-

change models and permeability models that couple different processes through a set of flow and

geomechanical equations and solve the system using numerical simulators. Gu and Chalaturnyk

(2005) and Palmer (2007), further divided analytical permeability-change models into either

stress or strain based models, with or without a geomechanical framework. Strain- based,

permeability-change models where defined as the ones where in the case of CBM production

desorption of methane changes the volumetric strain that causes the porosity to change and

finally affects the permeability. In the case of the stress-based models for CBM production,

desorption of methane changes the volumetric strain, which consequently changes the horizontal

stress and ultimately alters the permeability.

Settari and Walters (2001), Rutqvist et al (2002) and Liu (2011) divided permeability-

change models that couple fluid flow and solid deformation into decoupled simulation,

sequential coupled simulations and fully coupled simulations. In fully coupled simulations, both

reservoir flow variables and geomechanical variables are concurrently determined by solving a

system of equations. In sequential coupled simulation, parameters pertinent to flow, such as

pressure and temperatures, and geomechanical variables, such as stresses and displacements, are

solved sequentially first from a reservoir simulator and then from a geomechanical simulator and

coupling parameters, such as permeability and porosity are iterated between the two simulators

(Figure 3). In decoupled or uncoupled simulation the changes of fluid pressures cause the

changes of stresses and strains, but the changes of stresses and strains are assumed not to affect

fluid pressures (Wang, 2000).

Page 24: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

10

Figure 3 - Different Modes of Coupled Simulation (a, b, c) and Uncoupled Simulation (d)

Liu (2011), classified permeability-change models according to their reasoning of coalbed

configuration and assumptions postulated. More specifically, Liu considered a class of

permeability models that represents each coalbed as either a stack of matchsticks or as cubes

under uniaxial strain conditions (Figure 4) and constant overburden stress exerted by the

overlying geology. Liu (2011) also discussed another group of permeability models, where the

coalbed is represented by a stack of separated coal matrix matchsticks that are connected through

solid rock bridges of coal under reservoir conditions of variable stress state. Finally, several other

classifications can be postulated among models that account for isotropic or anisotropic

permeability changes within the coalbed.

Page 25: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

11

Figure 4 - Uniaxial Strain Model

Finally, Liu (2011) proposed upper and lower boundaries for the permeability-change

models. More specifically, the minimum permeability change is given under conditions of free

swelling and shrinkage as shown in Eq. (8) where the fracture aperture does not change with the

volumetric strain experienced by the coal matrix blocks since the boundary is free to move with

strain changes. In this case there is an increase in cleat spacing but the fracture aperture remains

constant. In comparison, the maximum permeability change is expected to occur as shown in Eq.

(9), under the constant volumetric model assumption where the fracture opening compensates the

volume changes in the coal matrix block (Figure 5).

[

(

)]

(8)

[

( )]

(9)

Figure 5 - Fracture Aperture Changes with Swelling/Shrinkage

Representative permeability-change models are shown below.

Vertical stress σzz

εxx=0

εzz<>0

εxx=0

εyy=0

Confining formation

Coal

Coal

Confining Formation

Swelling / Shrinkage

Page 26: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

12

Seidle at al., (1992), Pan et al., (2010), proposed a stress based model for permeability

changes, as shown in Eq. (10). In accordance with experimental results, an exponential function

was used for the permeability-stress relationship (Durucan and Edwards, 1986; Gray 1987;

Wang 2009). Further assumptions applied for this model, were the postulation of a matchstick

geometry for the coal matrix, isotropic permeability, uniaxial strain and constant overburden

stress.

[ ( )] (10)

where cf is the fracture cleat compressibility defined as follows

(11)

This model accounts for grain compaction and the cleats volume change contribution to the

permeability alteration however it does not incorporate volumetric strain due to gas adsorption or

desorption.

Gilman and Beckie (2000) suggested a stress-based relationship for isotropic permeability

under uniaxial strain and constant overburden stress conditions for saturated coals, as shown by

Eq. (12). An exponential function was employed to relate permeability to the other parameters in

the framework of poro-elasticity. A relative regular cleat system was assumed and it was also

considered that each fracture reacts as an elastic body with respect to change in the normal stress

component. In addition, an extremely slow mechanism of methane release from the coal matrix

to the cleats was postulated.

(

) [

(

)] (12)

Seidle and Huitt (1995) described a strain based permeability change model for isotropic

permeability of saturated coals in the framework of poro-elasticity under uniaxial strain and

constant overburden stress reservoir conditions, as shown in Eq. (13). A cubic function was

employed in this case to correlate the pertinent parameters to permeability. It was assumed that

Page 27: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

13

the coal sorption-induced strain is proportional to the amount of gas sorbed and that the sorbed

gas is related to pressure by Langmuir’s equation (Liu, 2011).

[

(

) (

)]

(13)

Shi and Durucan (2004) presented a stress based permeability model for isotropic

permeability changes in the framework of linear elasticity for saturated coals, as shown in Eq.

(14). An exponential function was employed to relate permeability to cleat compression and

matrix shrinkage. These two terms affect the dynamic permeability evolution in a competitive

way. For this model as in previous cases, uniaxial strain and constant overburden stresses were

assumed.

[ (

( )( )

( )

(

))] (14)

Palmer and Mansoori (1996), proposed a strain based, fully geomechanical model for

isotropic permeability under conditions of uniaxial strain and constant overburden stress in the

framework of linear elasticity for small strain changes, Eq. (15).

[

( ) (

) (

)]

( )( )

(15)

In 2007, Palmer and Mansoori modified their model to account for directional (anisotropic)

permeability and modulus changes with depletion. The Palmer and Mansoori (2007)

permeability change model is also applicable to both saturated and undersaturated coals. The

Computer Modelling Group (CMG) implemented this model in their commercial reservoir

simulator, GEM.

Peekot and Reeves (2002), suggested a permeability model in which matrix strain changes

are extracted from a Langmuir curve of strain versus reservoir pressure which is assumed to be

proportional to the adsorbed gas concentration curve (Palmer, 2007; Liu, 2011). This model is

not developed within a geomechanical framework. The matrix shrinkage is proportional to the

adsorbed gas concentration change multiplied by shrinkage compressibility ( ) that is a free

Page 28: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

14

parameter. This model is incorporated in the commercial reservoir simulator package, COMET3

by ARI (Comet, 2003). It has been shown that for saturated coals the latter model and the Palmer

and Mansoori (1996) model are essentially equivalent (Palmer, 2007; Liu 2011).

Cui and Bustin (2005), described a stress dependent permeability model for isotropic

permeability in the poro-elasticity framework under uniaxial strain and constant overburden

stress reservoir conditions, as described by Eq. (16). This model essentially accounts for the

effects of pressure and sorption induced volumetric strain on permeability constrained by the

adsorption isotherm.

{

[ ( )

( )( )

( )( )]} (16)

Gu and Chalaturnyk in 2010 proposed an anisotropic permeability-change model. As shown

in Eq. (17), they determined a directional effective strain, comprised of the directional

mechanical deformation due to stress change, the directional mechanical deformation due to

pressure change, the directional matrix shrinkage/swelling due to desorption/sorption and the

directional thermal contract/expansion due to temperature changes. The directional effective

strain was then related to a directional permeability, Eq. (18), with a cubic function and with

assuming a stack of matchsticks configuration for the coalbed.

(17)

(

)

(18)

Pan and Connell (2007) and Clarkson et al. (2008), proposed an elastic strain dependent

model for isotropic permeability under uniaxial strain and constant overburden stress conditions,

Eq. (19). This model accounts for sorption strain, Eq. (20), for a single component adsorption.

[

( )

(

) ]

( )( )

(19)

( ) ( )

( ) (20)

Robertson and Christiansen (2006) suggested an isotropic permeability-change model

assuming a cubic configuration for the coalbed under variable stress reservoir conditions and

Page 29: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

15

constant overlying geology, as shown in Eq. (21). In this model, effective porosity of the coal

matrix is considered to be equal to zero and thus only the porosity of the coalbed is attributed to

the fracture system.

{

[ ( )]

[

( )

(

)

]}

(21)

Zhang et al. (2008) presented a permeability change model under variable stress conditions

in the poro-elasticity framework, Eq. (22).

(

[( ) ( ])

(22)

where S and

(23)

(24)

Connell et al. (2010), Liu et al. (2010) and Liu and Rutqvist (2010), reported that when the

coalbed is represented by a stack of coal matrix matchsticks separated by the cleat system, under

conditions of constant confining stress (as is usually applied in laboratory experiments), then the

effects of coal matrix swelling will not alter the coal permeability. This statement argues that

since for a given pressure, the coal matrix blocks are fully separated by the fractures and no fixed

boundaries are enforced, the swelling will translate in an increase in fracture spacing and not in a

reduction of fracture aperture. Nevertheless, this conclusion does not tie in with laboratory

observations, where matrix swelling causes significant changes in coal permeability under

conditions of constant confining stress. As discussed by Liu (2010a), in order to account for this

interaction many researchers postulated a zero lateral strain condition in the horizontal plane so

that matrix swelling will result in permeability changes. (e.g., Harpalani and Chen, 1995;

Robertson and Christiansen, 2006; Harpalani and Chen, 1997; Pan et al., 2010; Pini et al., 2009).

Liu et al. (2010) suggested a different approach in order to correctly reproduce the

aforementioned laboratory observations, which accounts for interactions among coal matrix and

Page 30: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

16

fractures. More specifically, the cleat system in this model is not considered to be completely

separate neighboring matrix blocks, but assumes that there are solid bridges connecting the

adjacent coal matrix blocks (Figure 6). For this coalbed configuration, coal matrix-swelling

affects permeability in two opposing ways. When the bridging contacts swell, the coal matrix

blocks are driven apart and thus fracture aperture increases and consequently permeability. At

the same time, swelling of the matrix blocks results in reduction of fracture aperture and thus

permeability.

Figure 6 - Matchstick Model with Bridging Between Coal Blocks

Connell and Detournay (2009) suggested a sequential coupled simulation model to describe

dynamic evolution of permeability in an isotropic linear elasticity framework. Their formulation

was based on Shi and Durucan (2004) isotropic permeability-change stress based model.

However, they discussed a modified version in which directional permeability, triaxial strain and

varying overburden stress conditions were accounted for. Permeability changes for both CBM

and ECBM productions cases were examined by employing the SIMED reservoir simulator

(Spencer et al., 1987; Stevenson and Pinczewski, 1995) and the FLAC3D finite difference

geomechanical code. It was shown that especially for the ECBM production cases, simplifying

assumptions such as uniaxial strain and constant overburden stress can lead to large errors.

Gu and Chalaturnyk, 2010 proposed a sequential coupling permeability change model,

where the discontinuous coal masses were considered as an equivalent elastic continuum. Their

explicit-sequential coupled simulation examined pressure depleting CBM reservoirs. GEM

(CMG, 2003) a multidimensional, multiphase, isothermal and compositional reservoir simulator

and the geomechanical code FLAC3D (Itasca, 2002) designed for rock and soil mechanics

analyses were employed to model fluid flow and calculate coalbed deformation respectively. The

simulation process was as follows: (i) Calculate pore pressures, adsorbed gas volumes, water

saturation and well production rates of gas and water with GEM at an arbitrary time step, then

Coal

Bridging

Confining Formation

Page 31: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

17

(ii) insert pore pressures and adsorbed gas volumes in FLAC3D and calculate stresses and linear

strains and finally (iii) calculate cleat permeability based on Eq. (25).

(

)

(25)

Discussion on Implementation of Permeability Models for Reservoir

Simulation

Most commercially and in-house developed software for simulating reservoir behavior for

CBM/ECBM production has implemented one or more of the aforementioned analytical

approaches. In order to apply these relationships to specific geological strata sequences with

varying initial and boundary conditions, a numerical approach should be implemented. This

implementation requires discretization of the domain in two or three dimensions (Figure 7) and

the solution of the governing differential equations in space and time. Commonly finite

difference schemes are used to solve the flow problem. In the more advanced, complicated

approaches where the geomechanical component should also be taken into account a coupled

approach should be used. This can be accomplished either using external coupling (i.e. coupling

to software not directly interfaced with the flow package) or internal coupling where flow

equations can be modified directly by the stress strain regime of the reservoir as prescribed by

the analytical model (see also Figure 3).

The analysis provided in this study will be applied for the simulation of CBM and ECBM

production for a small-scale carbon capture, utilization and storage (CCUS) project in southwest

Virginia. More specifically, the injection and storage potential of unmineable coal seams will be

tested by injecting 20,000 tonnes of carbon dioxide into three wells, in Buchanan County,

Virginia, in the central Appalachian basin, over a period of one year (Figure 8).

Page 32: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

18

Figure 7 - Domain Discretization

Figure 8 - Case Study Area Showing Injection Wells

To successfully model reservoir response to simultaneous CO2 injection and CH4 production

several options are currently being examined. These courses of action include the following: a)

to use a commercially available package with a built in geomechanics component such as

GEM/CMG, b) to develop specific solutions for stresses and strains using an external package

such as FLAC3D, coupled to a commercially available flow simulator (Figure 9) and c) to use

other existing simulators developed by other research institutions with external or internal

coupling to geomechanics.

Page 33: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

19

Figure 9 - Initial Attempts for Coupled Simulation Using an External Geomechanics

Package (Vasilikou et al., 2012)

Conclusions

Coal permeability and permeability change models are probably the most important

considerations when evaluating the potential of a coal bed methane reservoir for enhanced coal

bed methane production. Different analytical models may be applied depending on the available

knowledge and properties of a reservoir. In the simplest considerations, the flow model will be

solved without input from the changing geomechanics regime in the area. However, as

complexity increases, other conditions should be taken into account (e.g., pore pressure changes,

shrinkage, swelling, etc.) and thus porosity and permeability fluctuations need to be incorporated

as dictated by changes in the stress and strain situation. Furthermore, in ECBM reservoirs

additional complications arise due to the presence of two gases in two-phase flow conditions.

In every step of conducting a reservoir simulation, the user should be aware of the net

contribution of each factor in the final result. Components with minimal contribution may be

safely eliminated after consideration. Also, the underlying assumptions as well as application

conditions should be evident for every analytical model, especially when performing history

matching. As already discussed in the literature, improper model hypothesis and model selection

will lead to large discrepancies between predicted and measured data.

The analytical model evaluation presented in this study will be utilized in reservoir

simulations for ECBM production for a small-scale carbon capture, utilization and storage

(CCUS) project in the central Appalachian basin.

Page 34: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

20

Acknowledgements

Financial assistance for this work was provided the U.S. Department of Energy through the

National Energy Technology Laboratory's Program under Contract No. DE-FE0006827.

Nomenclature

a the width of the coal

matrix block

Δb fracture aperture change

B Langmuir pressure

constant (Pa-1

)

Δp the change in pore pressure

b fracture aperture Δps equivalent sorption

pressure

bm the mechanical aperture

of a fracture

ΔS the change of the adsorbate

mass

cf coal cleat

compressibility

ΔT the change in reservoir

temperature

Cm shrinkage

compressibility

Δεe the change of total effective

volumetric strain

de the effective diameter of

grains

Δεeh the change of the effective

horizontal strain

E Young's modulus of coal ΔεlDi matrix shrinkage/swelling

due to desorption/sorption

Ef Young's modulus for the

fracture

ΔεlEi the mechanical deformation

due to stress change

Es Young's modulus of the

solid phase

Δεli the directional effective

strain

k coal permeability ΔεlPi the mechanical deformation

due to pressure change

K the bulk modulus of coal ΔεlTi thermal contract/expansion

due to temperatures changes

ki current permeability in i

direction

Δεs the change in sorption

strain

km coal matrix permeability Δεv the change of volumetric

strain of coal

km0 initial coal matrix

permeability

εe the total effective

volumetric strain

Kp the bulk modulus of coal

pore system

εL the maximum volumetric

strain

Ks the bulk modulus of coal

grains

εs the sorption strain for coal

L Langmuir sorption

constant (mol/kg)

εs0 the initial sorption strain of

coal

Page 35: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

21

p the gas pressure within

the pores

εv the volumetric strain of

coal

p0 the initial gas pressure

within the pores

ν Poisson's ratio of coal

PL Langmuir pressure

constant

νs Poisson's ratio for solid

phase

R the gas constant (8.314 J

mol− 1

K− 1

)

ρs the density for the solid

phase

s cleat spacing σh the horizontal stress

T reservoir temperature σh0 the initial horizontal stress

αc change rate in fracture

compressibility

ϕ coal porosity

αT the coefficient of

volumetric thermal

expansion

ϕ0 initial coal porosity

γ the volumetric

swelling/shrinkage

coefficient

ϕf the cleat porosity

References

Advanced Resources International, I. (2003). COMET3.

Avraam, D. G., & Payatakes, A. C. (1995). Flow regimes and relative permeabilities during

steady-state two-phase flow in porous media. Journal of Fluid Mechanics Digital Archive,

293, 207-236.

Ayers, W. B. (2002). Coalbed gas systems, resources, and production and a review of contrasting

cases from the San Juan and Powder River basins. AAPG Bulletin, 86, 1853–1890.

Bodden, W. P., & Ehrlich, R. (1998). Permeability of coals and characteristics of desorption

tests: implications for coalbed methane production. International Journal of Coal Geology,

35, 333–347.

Chen, Z., Liu, J., Connell, L., Pan, Z., & Zhou, L. (2008). Impact of Effective Stress and CH4–

CO2 Counter-Diffusion on CO2 Enhanced Coalbed Methane Recovery. Paper presented at

the SPE Asia Pacific Oil and Gas Conference and Exhibition, Perth, Australia.

Chilingar, G. V. (1964). Relationship between porosity, permeability, and grain-size distribution

of sands and sandstones. Paper presented at the 6th International Sedimentological

Congress.

Page 36: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

22

Clarkson, C. R., Pan, Z., Palmer, I. D., & Harpalani, S. (2008). Predicting Sorption-Induced

Strain and Permeability Increase With Depletion for CBM Reservoirs. Paper presented at the

SPE Annual Technical Conference and Exhibition, Denver, Colorado, USA.

Computer Modelling Group Ltd (CMG), C. (2003). GEM. Calgary, Canada.

Connell, L. D., Lu, M., Pan, Z. (2010). An analytical coal permeability model for tri-axial strain

and stress conditions. International Journal of Coal Geology, 84(2), 103-114.

Connell, L. D., & Detournay, C. (2009). Coupled flow and geomechanical processes during

enhanced coal seam methane recovery through CO2 sequestration. International Journal of

Coal Geology, 77(1-2), 222-233. doi: 10.1016/j.coal.2008.09.013

Cui, X., & Bustin, R. M. (2005). Volumetric strain associated with methane desorption and its

impact on coalbed gas production from deep coal seams. AAPG Bulletin, 89(9), 1181-1202.

Damen, K., Faaij, A., Van Bergen, F., Gale, J., & Lysen, E. (2005). Identification of early

opportunities for CO2 sequestration — worldwide screening for CO2-EOR and CO2- ECBM

projects. Energy, 30(10), 1931-1952.

Durucan, S., & Edwards, J. S. (1986). The effects of stress and fracturing on permeability of

coal. Mining Science & Technology, 3, 205-216.

Gash, B. W., Richard, F. V., Potter, G., & Corgan, J. M. (1992). The Effects of Cleat Orientation

and Confining Pressure on Cleat Porosity, Permeability and Relative Permeability in Coal.

Paper presented at the SPWLA/SCA Symposium, Oklahoma City.

Gilman, A., & Beckie, R. (2000). Flow of coal-bed methane to a gallery. Transport in Porous

Media, 41(1), 1-16.

Gray, I. (1987). Reservoir engineering in coal seams: part 1—the physical process of gas storage

and movement in coal seams. SPE Reservoir Engineering, 2(1), 28-34.

Gu, F. (2009). Reservoir and Geomechanical Coupled Simulation of CO2 Sequestration and

Enhanced Coalbed Methane Recovery. Doctor of Philosophy, University of Alberta,

Edmonton, Alberta.

Gu, F., & Chalaturnyk, J. J. (2005). Analysis of coalbed methane production by reservoir and

geomechanical coupling simulation. Journal of Canadian Petroleum Technology, 44(10).

Gu, F., & Chalaturnyk, R. (2010). Permeability and porosity models considering anisotropy and

discontinuity of coalbeds and application in coupled simulation. Journal of Petroleum

Science and Engineering 74(3-4), 113-131.

Page 37: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

23

Hamelinck, C. N., Faaij, A. P. C., Turkenburg, W. C., Van Bergen, H. J. M., Pagnier, Barzandji,

O. H. M., Wolf, K.-H., A., A., & Ruijg, G. J. (2002). CO2 enhanced coalbed methane

production in the Netherlands. Energy, 27, 647-674.

Harpalani, S., & Chen, G. (1997). Influence of gas production induced volumetric strain on

permeability of coal. Geotechnical and Geological Engineering, 15(4), 303-325.

Harpalani, S., & Chen, G. (1995). Estimation of changes in fracture porosity of coal with gas

emission. Fuel, 74(10), 1491-1498.

Harpalani, S., & Schraufnagel, R. A. (1990). Shrinkage of coal matrix with release of gas and its

impact on permeability of coal. Fuel, 69(5).

Itasca. (2005). FLAC3D: Fast Lagrangian analysis of continua in 3 dimensions. Theory and

background. Minnesota, USA.

Law, D. H. S., Meer, v. d., L.G.H., & Gunter, W. D. (2002). Numerical simulation comparison

study for enhanced coalbed methane recovery processes. Paper presented at the Gas

Technology Symposium Calgary.

Liu, F., Lu, P., Zhu, C., & Xiao, Y. (2010a). Coupled reactive flow and transport modeling of

CO2 sequestration in the Mt. Simon sandstone formation, Midwest U.S.A. International

Journal of Greenhouse Gas Control. doi: doi:10.1016/j.ijggc.2010.08.008

Liu, H.-H., & Rutqvist, J. (2009). A New Coal-Permeability Model: Internal Swelling Stress and

Fracture–Matrix Interaction. Transport in Porous Media, 82(1), 157-171. doi:

10.1007/s11242-009-9442-x

Liu, J., Chen, Z., Elsworth, D., Miao, X., & Mao, X. (2010b). Evaluation of stress-controlled

coal swelling processes. International Journal of Coal Geology, 83(4), 446-455.

Liu, J., Chen, Z., Elsworth, D., Miao, X., & Mao, X. (2010c). Linking gas-sorption induced

changes in coal permeability to directional strains through a modulus reduction ratio.

International Journal of Coal Geology, 83(1), 21-30.

Liu, J., Chen, Z., Elsworth, D., Qu, H., & Chen, D. (2011). Interactions of multiple processes

during CBM extraction: A critical review. International Journal of Coal Geology, 87(3-4),

175-189. doi: 10.1016/j.coal.2011.06.004

Liu, J., Chen, Z., & Qu, H. (2010d). Multiphysics of coal–gas interactions: the scientific

foundation for CBM extraction. Paper presented at the SPE Asia Pacific Oil and Gas

Conference and Exhibition, Queensland, Australia.

Page 38: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

24

Lu, M., & Connell, L. D. (2007). A model for the flow of gas mixtures in adsorption dominated

dual porosity reservoirs incorporating multi-component matrix diffusion: Part I. Theoretical

development. Journal of Petroleum Science and Engineering, 59(1-2), 17-26.

Marsh, H. (1987). Adsorption methods to study microporosity in coals and carbons—a critique.

Carbon, 25(49-58).

Mavor, M. J., Gunter, W. D., Robinson, J. R., Law, D. H. S., & Gale, J. (2002). Testing for CO2

Sequestration and Enhanced Methane Production From Coal. Paper presented at the SPE

Gas Technology Symposium, Richardson, Texas, USA.

Palmer, I. D., Mavor, M., Gunter, B. (2007). Permeability changes in coal seams during

production and injection. Paper presented at the International Coalbed Methane

Symposium., Tuscaloosa, Alabama.

Palmer, I. D., & Mansoori, J. (1996). How permeability depends on stress and pore pressure in

coalbeds: a new model. Paper presented at the SPE Annual Technical Conference and

Exhibition, Denver, Colorado.

Pan, Z., & Connell, L. D. (2007). A theoretical model for gas adsorption-induced coal swelling.

International Journal of Coal Geology, 69(4), 243-252.

Pan, Z., Connell, L. D., & Camilleri, M. (2010). Laboratory characterisation of coal reservoir

permeability for primary and enhanced coalbed methane recovery. International Journal of

Coal Geology, 82(3-4), 257-267.

Patching, T. H. (1970). Retention and release of gas in coal — a review. Canadian Mining and

Metallurgical Bulletin, 63(703), 1302-1308.

Peekot, L. J., & Reeves, S. R. (2002). Modeling the Effects of Matrix Shrinkage and Differential

Swelling on Coalbed Methane Recovery and Carbon Sequestration: U.S. Department of

Energy.

Pini, R., Ottiger, S., Burlini, L., Storti, G., & Mazzotti, M. (2009). Role of adsorption and

swelling on the dynamics of gas injection in coal. Journal of Geophysical Research,

114(B4), B04203.

Puri, R., & Yee, D. (1990). Enhanced Coalbed Methane Recovery. Paper presented at the SPE

Annual Technical Conference and Exhibition, New Orleans, Louisiana.

Reiss, L. H. (1980). The Reservoir Engineering Aspects of Fractured Formations, Houston.

Page 39: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

25

Robertson, E. P. (2005). Modeling permeability in coal using sorption-induced strain data. Paper

presented at the SPE Annual Technical Conference and Exhibition, Dallas, Texas.

Robertson, E. P., Christiansen. (2006). A permeability model for coal and other fractured,

sorptive-elastic media. Paper presented at the SPE Eastern Regional Meeting, Canton, Ohio,

USA.

Rutqvist, J., Wu, Y.-S., Tsang, C.-F., & Bodvarsson, G. (2002). A modeling approach for

analysis of coupled multiphase fluid flow, heat transfer, and deformation in fractured porous

rock. Int. J. Rock Mech. Min. Sci., 39, 429-442.

Seidle, J. P., Jeansonne, M. W., & Erickson, D. J. (1992). Application of matchstick geometry to

stress dependent permeability in coals. Paper presented at the SPE Rocky Mountain

Regional Meeting, Wyoming.

Seidle, J. R., & Huitt, L. G. (1995). Experimental measurement of coal matrix shrinkage due to

gas desorption and implications for cleat permeability increases. Paper presented at the

International Meeting on Petroleum Engineering, Beijing, China.

Settari, A., Walters, D. A., & Duke Engineering and Services Inc, D. E. a. S. (2001). Advances

in Coupled Geomechanical and Reservoir Modeling With Applications to Reservoir

Compaction. SPE Journal, 6(3), 334-342.

Shi, J. Q., & Durucan, S. (2004). Drawdown induced changes in permeability of coalbeds: a new

interpretation of the reservoir response to primary recovery. Transport in Porous Media

56(1), 1-16.

Shi, J. Q., & Durucan, S. (2005). CO2 Storage in Deep Unminable Coal Seams. Oil & Gas

Science and Technology – Rev. IFP, 60(3).

Siriwardane, H., Haljasmaa, I., McLendon, R., Irdi, G., Soong, Y., & Bromhal, G. (2009).

Influence of carbon dioxide on coal permeability determined by pressure transient methods.

International Journal of Coal Geology, 77(1-2), 109-118. doi: 10.1016/j.coal.2008.08.006

Spencer, S. J., Somers, M.L., Pinczewski, W.V., Doig, I.D. (1987). Numerical simulation of gas

drainage from coal seams. Paper presented at the 62nd SPE Annual Technical Conference

and Exhibition, Houston.

Stanton, R., Flores, R., Warwick, P. D., Gluskoter, H., & G.D., S. (2001). Coalbed Sequestration

of Carbon Dioxide. Paper presented at the 1st National Conference on Carbon Sequestration,

Washington, USA.

Page 40: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

26

Stevenson, M., & Pinczewski, V. (1995). SIMED II — Multi-component Coalbed Gas

Simulator. User's Manual Version 1.21: Centre for Petroleum Engineering, University of

New South Wales.

Vasilikou, F., Keles, C., Ripepi, N., & Gilliland, E. (2012). Coupling Reservoir and

Geomechanical Models to Assess Enhanced Coalbed Methane Recovery in Central

Appalachian Coal. Pittsburgh: US DOE/NETL.

Wang, G. X., Massarotto, P., & Rudolph, V. (2009). An improved permeability model of coal for

coalbed methane recovery and CO2 geosequestration. International Journal of Coal Geology,

77(1-2), 127-136. doi: 10.1016/j.coal.2008.10.007.

Wang, H. F. (2000). Theory of Linear Poroelasticity with Application to Geomechanics and

Hydrogeology. USA: Princeton University Press.

Warren, J. E., Root, P. J., & CO., G. R. A. D. (1963). The Behavior of Naturally Fractured

Reservoirs. SPE Journal, 3(3), 245 - 255.

Zarrouk, S. J., & Moore, T. A. (2009). Preliminary reservoir model of enhanced coalbed methane

(ECBM) in a subbituminous coal seam, Huntly Coalfield, New Zealand. International

Journal of Coal Geology, 77(1-2), 153-161. doi: 10.1016/j.coal.2008.08.007.

Zhang, H., Liu, J., & Elsworth, D. (2008). How sorption-induced matrix deformation affects gas

flow in coal seams: a new FE model. International Journal of Rock Mechanics and Mining

Sciences, 45(8), 1226-1236.

Page 41: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

27

EXPERIENCES IN RESERVOIR MODEL CALIBRATION FOR COAL

BED METHANE PRODUCTION IN DEEP COAL SEAMS IN RUSSELL

COUNTY, VIRGINIA2

Foteini Vasilikou, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Cigdem Keles, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Zach Agioutantis, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Nino Ripepi, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Michael Karmis, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Abstract

Injection and storage of carbon dioxide (CO2) on the coal seam matrix has two benefits:

mitigation of greenhouse gas emissions and enhanced recovery of coalbed methane (ECBM).

The theory of ECBM is based on the natural affinity for the porous coal matrix to preferentially

adsorb CO2 over methane. According to laboratory measurements on bituminous Appalachian

coals, the coal matrix can adsorb almost twice the amount of CO2 by volume as methane at

2 Experiences in Reservoir Model Calibration for Coal Bed Methane Production in Deep

Coal Seams in Russell County, Virginia.

F. Vasilikou, C. Keles, Z. Agioutantis, N. Ripepi, M. Karmis. 2013 Symposium on

Environmental Considerations in Energy Production, SME, April 14-18, Charleston, West

Virginia.

Reprinted with permission of SME.

Foteini Vasilikou researched and prepared this manuscript, with Cigdem Keles, Zach

Agioutantis, Nino Ripepi and Michael Karmis providing technical and editorial input.

Page 42: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

28

typical reservoir pressures. Based on the fact that coalbed methane (CBM) is currently

economically produced in many areas in the world, presents the likelihood that CO2 injected

under the appropriate conditions could sequester the CO2 while displacing and producing CBM.

The Coal Seam Group of the Southeast Regional Carbon Sequestration Partnership

(SECARB) has recently completed the injection of 1,000 tons of carbon dioxide into multiple

deep unminable coal seams as part of a field validation test at their Russell County, VA site in

Central Appalachia and is planning the injection of 20,000 tons of CO2 in a nearby field in

Buchanan County, VA. This paper presents the reservoir modeling procedure and the parameters

developed in order to obtain an accurate history match for both gas and water production for a

group of eight mature CBM wells that were utilized in the Russell County field test. Optimal

parameters were determined both for the single well models as well as for the combined eight-

well model. This model will form the base model for simulating CO2 injection and subsequently

predicting post-injection behavior for the upcoming injection test at the Buchanan county site.

Introduction

The Coal Seam Group of the Southeast Regional Carbon Sequestration Partnership

(SECARB) has recently completed the injection of 1,000 tons of carbon dioxide into multiple

deep unminable coal seams as part of a field validation test at their Russell County, VA site in

Central Appalachia (Ripepi et al, 2009).

Figure 10 shows an aerial view of the general area, while Figure 11 shows an aerial view of

the injection well (BD114), two monitoring wells (black dots) and the closest offset producing

CBM wells.

Page 43: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

29

Figure 10 - Location of Field Test Site (Ripepi et al, 2009)

Figure 11 - Layout of Injection and Offset Wells (Karmis et al., 2008)

Prior to conducting the injection test a reservoir simulator was built which predicted the

extent of the CO2 plume based on history matching of CH4 and water production data (VCCER,

2011). In this paper the original reservoir simulation is extended by using new knowledge

BD 114

BC 114

BD 113

BC 115

BD 115

BE115

BE114

BE113

1604'

14

33

'

2167 '

C ircle R adius1320'

1311'

21

69

'

16

36

'

2297 '

0' 1000 '500 '

Page 44: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

30

regarding seam geometry, coal properties (gas content, adsorption isotherms, anisotropy in

permeability), butt and face cleat orientation and by considering well interference.

Geology

A carbon dioxide injection site was selected in Russell County, Virginia, where an existing

CBM well was donated by CNX Gas Corporation, a subsidiary of CONSOL Energy Inc. The

injection well may be found referenced by its company well name, BD114, or by the State of

Virginia designation, RU-0084. BD114 is located at an elevation of 2,523 feet in the Honaker

District of Russell County, Virginia.

The Central Appalachian Basin is a northeast-to-southwest trending basin encompassing

approximately 10,000 square miles in southwestern Virginia, southern West Virginia and eastern

Kentucky (Conrad et al., 2006). Production of CBM began in 1988 with the development of the

Nora Field in Dickenson County, Virginia followed by CONSOL Energy developing the

Oakwood Field in Buchanan County, Virginia. Since that time, over 5,600 CBM wells have

been drilled and brought on-line as producing gas wells in southwest Virginia through 2012

(VaDMME, 2013). The coals in the region near the injection site include those of the

Pocahontas Formation and Lee Formation which directly overlies the late Mississippian

Bluestone Formation. Coal seams of the Pocahontas and Lee Formation are medium to low-

volatile bituminous, high rank and high gas content coals that include the Pocahontas No. 1

through Pocahontas No. 9 seams (Pocahontas Formation and the Upper Seaboard, Greasy Creek,

Middle and Lower Seaboard, Upper and Middle Horsepen, War Creek, and Lower Horsepen

coals (Lee Formation).

The Upper Horsepen (UH) 2&3 coal seams are the thickest single completion at 3.7 feet of

net coal at a depth of 1,374 feet, with a 0.2 feet parting between the seams. The Pocahontas No. 3

coal seam is 2,208 feet deep and 2.4 feet thick at the test site and well outside of current deep

mining activities.

The primary confining units include multiple layers of low permeability shale and siltstone

beds that range in thickness from five to 55 feet in the vicinity of the injection well. Permeability

for the shale and siltstone units is expected to range from 0.001 to 0.1 md, with low porosity.

Even the sandstone units are considered to provide confining beds due to the well cemented

nature of these rocks. The Lee and Pocahontas Formation sandstones are known to have low

Page 45: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

31

permeability and porosity values, and do not produce natural gas in this area. The tight sandstone

units range in thickness from five to 60 feet at the injection site. The sandstone units are expected

to have porosity values that range from 1.0 to 3.0 percent and permeability values ranging from

0.1 to 1.0 md based on core testing. All lithologies will encounter some natural fractures;

however, these fractures are likely to be cemented with quartz and calcite and are not expected to

provide permeability pathways based on core analysis, well logging and field testing (Karmis et

al., 2008).

Face and butt cleat planes were measured at two coal seam outcrop locations near the site

and they corresponded to within 7 degrees of the known face and butt cleat directions of N18W

and N167E respectively, for the deep mined Pocahontas No. 3 coal seam in Buchanan County. A

Rose Diagram (Figure 12) was developed to graphically display the cleat directions from the

injection well (VCCER, 2011).

Figure 12 - (A) Rose Diagram Showing Face and Butt Cleat Orientation (B) Rotation by

35o Clockwise

Well Stimulation

Twenty-four separate coal seams, totaling 36.3 feet, between the depths of 1,044 and 2,259

feet were perforated through the casing (Table 1) for well stimulation. The coal seams perforated

for the four stage well stimulation average 1.5 feet thick.

0

90

180

3

2

1

2

11

-114

- Average Butt Cleat Orientation: 74o

Average Face Cleat Orientation: 167o

Direction of Max Horizontal Stress: 55o

M1

M2

90o

180o

270o

360o

BD-114/RU-84

0

90

180

3

2

1

2

11

-114 -

Butt CleatsFace Cleats

Direction of Max Horizontal Stress

M1

M2

90 o

180 o

270 o

360 o

BD-114/RU-84

J Direction

I Direction

Page 46: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

32

BD114 was stimulated over perforations ranging in depth from 1,044.5 to 2,258 feet (318.36

to 688.24 m), using a four-stage nitrogen foam hydraulic fracturing treatment (VaDMME, 2002).

These four stages included 24 separate coal seams with a combined coal thickness of 36.3 feet

(11.1 m), which were completed (i.e. perforated and fracked) (Table 1). The first stage of the

injection well was not considered completed and was abandoned due to loss of energy and

“sanding off.” Therefore, a packer was set at 1,400 feet in the casing of BD114 and a string of 2

7/8-inch (7.2 cm) tubing was run back to the surface to isolate the injection zone. Thus, the

interval for injection comprised only the first three stages of the stimulation program.

Page 47: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

33

Table 1 - BD114 Stimulated Coal Seams (VaDMME, 2002)

Hydraulic

Fracture

Stage

Coal Seam Depth to

Coal

Coal

Thickness

Cumulative

Completed

Coal

Thickness

Zone

Thickness

(ft) (m) (ft) (m) (ft) (m) (ft) (m)

Stage 4

Greasy Creek 1 1,045 318.4 1.2 0.4 1.2 0.4

9.6 3.0

Seaboard 2 1,191 362.9 0.5 0.2 1.7 0.6

LowerSeaboard

1&2 1,251 381.3 2.0 0.6 3.7 1.2

Lower Seaboard 3 1,313 400.3 2.2 0.7 5.9 1.9

UpperHorsepen

2&3 1,374 418.9 3.7 1.1 9.6 3.0

Stage 3

Middle Horsepen 1 1,421 433.0 2.2 0.7 11.8 3.7

9.8 3.0

Middle Horsepen 2 1,497 456.4 0.7 0.2 12.5 3.9

Pocahontas 11 1,547 471.5 1.9 0.6 14.4 4.5

Pocahontas 10 1,573 479.5 1.0 0.3 15.4 4.8

Lower Horsepen 1 1,622 494.4 2.1 0.6 17.5 5.4

Lower Horsepen 2 1,627 495.9 1.9 0.6 19.4 6.0

Stage 2

Pocahontas 9 1,664 507.0 1.8 0.5 21.2 6.5

9.3 2.8

Pocahontas 8-1 1,710 521.2 2.0 0.6 23.2 7.1

Pocahontas 8-2 1,725 525.8 1.3 0.4 24.5 7.5

Pocahontas 7-1A 1,758 535.9 0.8 0.2 25.3 7.7

Pocahontas 7-1B 1,765 538.0 0.9 0.3 26.2 8.0

Pocahontas 7-2 1,850 563.9 1.6 0.5 27.8 8.5

Pocahontas 7-3 1,875 571.5 0.9 0.3 28.7 8.8

Stage 1

Pocahontas 6 1,984 604.8 0.5 0.2 29.2 9.0

7.6 2.3

Pocahontas 5 2,033 619.7 0.7 0.2 29.9 9.2

Pocahontas 4-1 2,125 647.7 1.9 0.6 31.8 9.8

Pocahontas 4-2 2,148 654.7 1.1 0.3 32.9 10.1

Pocahontas 3-1 2,208 673.0 2.4 0.7 35.3 10.8

Pocahontas 3-4 2,258 688.3 1.0 0.3 36.3 11.1

Total 24 Coal Seams 36.3 11.1 36.3 11.1 36.3 11.1

Page 48: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

34

Production History

BD114 was brought on-line as a CBM producing well in 2002. Water is produced through a

2 7/8-inch string of tubing set below the Pocahontas No. 3 coal seam, and gas is produced

between the casing and the tubing (annulus). Up until injection commenced in 2009, BD114 has

averaged production of 1.19 thousand cubic meters (Mm3) (42 thousand cubic feet (Mcf)) of

natural gas per day and 2.2 barrels of water per day (Figure 13). BD114 produced 2.89 MMm3

(102 MMcf), and 5,360 barrels of water prior to being taken off-line for conversion to an

injection well. Gas production to date is nearly 10 percent of the estimated gas in place (VCCER,

2011). BD114 is a below average gas producer for this gas field. The average production of the

seven offset wells is 1.87 Mm3 (66 Mcf) per day of natural gas and 2.5 barrels per day of water.

Figure 13 presents the gas production history for the injection well and the nearby offset wells

(VCCER, 2011).

Figure 13 - Gas Production History at the Injection and the Offset Wells

0

20

40

60

80

100

120

140

03/2

002

09/2

002

03/2

003

09/2

003

03/2

004

09/2

004

03/2

005

09/2

005

03/2

006

09/2

006

03/2

007

09/2

007

03/2

008

09/2

008

03/2

009

Ga

s P

rod

uc

tio

n (

Mc

f/d

ay

)

BD114 BE113 BE114 BC114 BD113 BC115 BD115 BE115

Page 49: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

35

Modeling

Description of Production Model

Preliminary reservoir modeling prior to the actual field test where CO2 was injected into the

coal seams was conducted utilizing Advanced Resources’ COMET3 reservoir simulator for

CBM. The ultimate goal of this preliminary modeling was to estimate the size of the CO2 plume

as well as the expected CO2 injection rate that could be achieved during injection operations. As

additional laboratory analysis, field data and results from the CO2 injection field test became

available the need to revise the original reservoir model became apparent. At the same time,

Computer Modeling Group Ltd.’s GEM software, a compositional and unconventional reservoir

simulator, was utilized since it could simulate the effect of fractures. This paper discusses the

setup and parameters for production modeling of the Russell county field test using CMG’s

reservoir simulator. The ultimate goal of the current study will be to obtain a good history match

that will help modeling of Enhanced Coal Bed Methane production after the injection well comes

back on line. The latter work is detailed in Vasilikou et al. (2013).

A dual porosity, single permeability model was setup, since coal matrix permeability is

considered extremely low. As mentioned before, there were four fracture stages including a total

of 24 coal seams. Gas and water rates were available per well and therefore, only aggregate

production was available for all seams. Since coal seams have different characteristics it was

decided to model four equivalent coal seams at the highest elevation of each fracture stage. The

thickness of these modeled coal seams corresponds to the sum of the thicknesses of each seam

within the fracture stage. A grid was generated that covers all wells with nearly the same

distance from the boundaries for all wells.

The orientation of the hydraulic fractures was critical in determining grid generation. As the

model can only allow for hydraulic fractures along either the I (x) or J (y) direction, and

hydraulic fractures were developed along the maximum horizontal stress in the area (Figure 12),

the grid was rotated by 35 degrees clockwise (-35o) to orient the maximum horizontal stress in

the I direction. In addition, it was assumed that the permeability in the I direction is equal to butt

cleat permeability and that in the J direction equal to face cleat permeability.

Page 50: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

36

Furthermore, it was assumed that the modeled group of eight wells can be considered

independent of other wells surrounding this grid. The resulting grid for all layers including the

eight wells is shown in 3D view in Figure 14.

Although, actual fracture modeling was not accomplished within the scope of this study,

nevertheless, the foundation was set and results will be presented in the upcoming paper by

Vasilikou et al. (2013).

Figure 14 - 3D View of Model Grid and Wells

The parameters and material properties shown in Table 2 were assembled utilizing data from

(a) collected from monitoring well BD114-M2 (Figure 12), (b) coal sample analysis (Harpalani,

2012).

In addition it should be noted that historical production data for gas and water were

aggregated on a monthly basis.

Page 51: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

37

Table 2 - Parameters and Material Properties Used in the Reservoir Model

Number of Grid Blocks per Direction (combined model) I= 84, J= 56

Number of Grid Blocks per Direction (single-well models) Variable (typical I=25,J=14)

Grid Block Sizes per Direction (combined model), m DI=25, DJ=25

Coal Layers and Thicknesses, m 4 Layers;(i) 2.93, (ii) 2.98, (iii) 2.83, (iv) 2.31

Coal Layers Depths- Grid Tops, m 318.4, 433, 507, 604.8

Coal (Matrix) Porosity, % 6.5%, dual porosity model

Cleat (Fracture) Porosity, % Variable, 1-3%, dual porosity model

Matrix Permeability, mD Not in use, single permeability model

Cleat Permeability, mD Variable (domain divided into 6 regions), 2-18.97 mD,

face to butt cleats ratio:1-5

Fracture Spacing per Direction, m I: 0.00508, J: 0.00508, K: 0.00508

Well Radius, m 0.088

Formation Compressibility

(Matrix & Fracture), 1/kPa

2.90075E-07 (at reference pressure 4136.8 kPa)

(Harpalani, 2012)

Reservoir Temperature, oC 27.2

Water Density, kg/m3 1000.8 at reference pressure 101.3 kPa

Relative Permeability Curves

Gas-Water

Gash et al. 1993, Mavor et al. 1953 (see Figure 15)

Adsorption Isotherms for Methane

according to Langmuir.

Langmuir Constants

VL, gmole/kg; PL, 1/kPa

Layer 1: Pocahontas 11,

VL= 0.8201299, PL= 0.0005224

Layer 2: Pocahontas 11,

VL= 0.8559385, PL= 0.0005224

Layer 3: Pocahontas 7,

VL= 0.8201299, PL= 0.00047818

Layer 4: Pocahontas 3,

VL= 0.7700043, PL= 0.00034816

(VCCER, 2011)

Coal Density, kg/m3 1323.9 (VCCER, 2011)

Methane Desorption Time, days 50

Pore Pressure Gradient, kPa/m 7

Page 52: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

38

Figure 15 - Relative Permeability Curves Used in the Single Well Runs (Gash et al. 1993,

Mavor et al. 1953) (SWT: Water Relative Permeability or Water Saturation Fraction;

SLT: Gas Relative Permeability or Gas Saturation Fraction) (see Table 2)

Previous work (VCCER, 2011) had shown that cleat permeability may be different around

each producing well. Under this assumption eight single-well models were setup which

comprised 25 to 50 cells in the X direction, approximately 20 cells in the Y direction and 1 cell

in the Z direction for each of the four modeled seams. Single-well models used the cleat

permeability regions shown in Figure 16 and Figure 17.

Page 53: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

39

Figure 16 - Face Cleat Permeability Regions for Single-Well Models and Combined Model

Figure 17 - Butt Cleat Permeability Regions for Single-Well Models and Combined Model

Page 54: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

40

Results and Discussion

Table 3 shows the optimal parameters determined for the single-well runs. The bottom hole

pressure (assumed to be equal to the reservoir pore pressure during production) also follows an

expected trend, i.e. increased for higher water production, and with a decreasing trend as gas rate

values increase. Note that these models do not include modeling of hydraulic fractures. Models

with hydraulic fractures are discussed in Vasilikou et al. (2013).

Table 3 - Optimal Parameters Determined for the Dingle-Well Runs

Well Fracture

Porosity

Face Cleat

Permeability

(mD)

Butt Cleat

Permeability

(mD)

Face to Butt

Cleat

Permeability

Ratio

Bottom

Hole

Pressure

(kPa)

Water

Saturation

in the Cleats

RU-84 0.0013 18.97 6.32 3 200 1

RU-85 0.001 14.42 4.12 3.5 700 1

RU-112 0.0013 14.42 4.12 3.5 1200 0.9

RU-123 0.001 14.42 4.12 3.5 200 0.8

RU-132 0.001 10.42 3.47 3 200 0.9

RU-210 0.001 10 7.8 1.28 1300 1

RU-211 0.001 10 2 5 1400 1

RU-284 0.001 7.8 7.8 1 1250 1

Figure 18 presents the reservoir simulation history-match results for BD-115 (RU-132) for a

single well run. Gas rate predictions match exactly the measured data, while water rate

predictions are adequate based on the quality of measured water data. The data for water

production by well are estimated by dividing the total water production of a group of wells by

the time the water pump on each well operated, unlike gas production which is measured at each

well. The bottom hole pressure (assumed to be equal to the reservoir pore pressure during

production) also follows an expected trend, i.e. increased for higher water production, and with a

decreasing trend as gas rate values increase.

Figure 19 presents the reservoir simulation history-match results for BD-115 (RU-132) for a

combined run where the single run optimum parameters were utilized as inputs for each well. In

both cases the gas rate match is perfect as the primary simulation constraint comprises matching

the gas rate.

Page 55: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

41

Figure 18 - History Match on Gas and Water Rates for Well BD-115 (RU-132) Under a

Single Well Run. The Bottom Hole Pressure is Also Shown

Figure 19 - History Match of Gas and Water Rates for Well BD-115 (RU-132) for the

Combined Run. The Bottom Hole Pressure is Also Shown

Page 56: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

42

The optimal parameters for the combined model were then determined using the

optimization algorithm (CMOST) built into the CMG/GEM package. The initial range for each

optimized parameter was determined using data ranges that were based on the data in Table 3 for

the single-well runs. The actual data ranges entered into the optimizer are shown in Table 4. The

optimization algorithm was instructed to find optimum parameter combinations in these ranges.

The objective function was set to minimize the error on cumulative water while the operating

constraint was to match the gas produced. The face to butt cleat permeability ratio was kept at 3

for all runs because preliminary runs showed that approximate ratio to be valid for all wells.

Although the algorithm should examine all possible combinations, some parameter combinations

are evidently not tested due to some internal optimization. In total 835 runs were completed

using CMOST. Table 5 shows the results obtained for all wells.

Table 4 - Parameter Range Used as Input in the CMOST Model Runs

Fracture

Porosity

Butt Cleat Permeability

(mD)

Bottom Hole

Pressure (kPa)

Water Saturation

in the Cleats

0.0010 1 200 0.900

0.0015 2 500 0.925

0.0020 3 800 0.950

0.0025 4 1100 0.975

0.0030 5 1400 1.000

6 1700

2000

In addition the optimum relative permeability curves were determined by the CMOST

algorithm. Two of these diagrams are shown in Figure 20. The fact that different optimum curves

are calculated for each well indicates that the relative permeability curves have a significant

impact on the gas and water rate calculations, may differ considerably between wells.

Page 57: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

43

Table 5 - Optimal Parameters After CMOST Runs and Manual Adjustments for the Multi-

Well Runs

Well Fracture

Porosity

Face Cleat

Permeability

(mD)

Butt Cleat

Permeability

(mD)

Face to Butt

Cleat

Permeability

Ratio

Bottom

Hole

Pressure

(kPa)

Water

Saturation

in the Cleats

RU-84 0.0020 12.0 4.0 3 200 0.950

RU-85 0.0011 10.5 3.5 3 200 0.975

RU-112 0.0010 9.0 3.0 3 800 0.950

RU-123 0.0010 18.0 6.0 3 500 0.800

RU-132 0.0026 16.8 5.6 3 1100 0.975

RU-210 0.0015 9.0 3.0 3 1500 0.950

RU-211 0.0025 9.0 3.0 3 1700 0.950

RU-284 0.0010 9.0 3.0 3 800 0.900

Figure 20 - Relative Permeability Curves as Determined by the Optimization Algorithm for

Two of the Eight Production Wells (SWT: Water Relative Permeability or Water

Saturation Fraction; SLT: Gas Relative Permeability or Gas Saturation Fraction)

Page 58: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

44

Figure 21 - History Match on Gas and Water Rates for Well BD-115 (RU-132) for the

Combined Run with Optimized Input Parameters. The Bottom Hole Pressure is Also

Shown

Figure 21 presents the reservoir simulation history-match results for BD-115 (RU-132) for a

combined run where the optimal parameters from the combined well optimization were used. It

is clear that the calculated water rate production matches the measured data better than in the

case of the previous two models (single or combined run). At the same time the maximum pore

pressure calculated for this production history match is higher than in previous runs, reaching

almost 3500 kPa at its maximum.

Conclusions

In this paper, the modeling procedure for coal bed methane production of selected wells in

Russell County, VA is discussed in detail. Through parametric analysis and optimization, the

best parameters for this reservoir are identified. Utilizing these parameters, the model can

Page 59: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

45

accurately predict the gas production rate for all wells through history matching, while the total

water produced from the reservoir matches reasonably well the measured values.

In addition, the model identifies the most sensitive parameters in the simulation and how

these interact in closely spaced wells. The development of this model along with the optimal

parameters that were calculated will be subsequently used for modeling of CO2 injection and the

corresponding enhanced gas recovery potential in the same group of seams.

Acknowledgments

Financial assistance for this work was provided by the U.S. Department of Energy through

the National Energy Technology Laboratory’s Program under Contract No. DE-FC26-

04NT42590 and DE-FE0006827.

References

Gash, B.W., Volz, R.F., Potter, G., and Corgan, J. M. 1993. The effects of cleat orientation and

confining pressure on cleat porosity, permeability and relative permeability in coal. Paper

9321 in Proceedings of the 1993 International CoalBed Methane Symposium. Tuscaloosa:

University of Alabama.

Mavor, M.J. and Robinson, J.R., 1993, “Analysis of Coal Gas Reservoir Interference and Cavity

Well,” Paper SPE, 25860, presented at the Joint Rocky Mountain Regional and Low

Permeability Reservoirs Symposium, Denver, CO, April 26-28.

Computer Modelling Group Ltd (CMG), C. (2003). GEM. Calgary, Canada.

Ripepi, N., Karmis, M., Miskovic, I., Shea C., and J.M. Conrad. 2009. Results from the Central

Appalachian Basin Field Verification Test in Coal Seams. 26th Annual International

Pittsburgh Coal Conference, Pittsburgh, PA, USA, 2009. CD-ROM of Proceedings, Paper #

33-4.

Karmis, M., Ripepi, N., Miskovic, I., Conrad, M., Miller, M., and C. Shea. 2008. CO2

Sequestration in Unminable Coal Seams: Characterization, Modeling, Assessment and

Testing of Sinks in Central Appalachia. Twenty-Fifth Annual International Pittsburgh Coal

Conference, Pittsburgh, PA, USA, 2008. CD-ROM of Proceedings, Paper # 29-3.

Page 60: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

46

Conrad, J. M., Miller, M. J., Phillips, J., Ripepi, N. 2006. Characterization of Central

Appalachian Basin CBM Development: Potential for Carbon Sequestration and Enhanced

CBM Recovery, 2006 International Coalbed Methane Symposium, Preprint 0625,

Tuscaloosa, AL.

Virginia Department of Mines, Minerals and Energy (VaDMME). 2013.

http://www.dmme.virginia.gov/DGO/Production/2010-County.pdf, accessed: February

2013.

Conrad, J. M. , Miller, M. J., Ripepi N. 2007. Potential for Carbon Sequestration in the Central

Appalachian Basin. Air and Waste Management Association Conference, 2007, Paper # 187.

Vasilikou, F. C. Keles, Z. Agioutantis, N. Ripepi and M. Karmis, 2013. Model Verification of

Carbon Dioxide Sequestration in Unminable Coal Seams with Enhanced Coal Bed Methane

Recovery, to be published in 23rd World Mining Congress, August 11-15, Montreal,

Canada.

Virginia Center for Coal and Energy Research (VCCER). 2011. Final Technical Report:

Characterization and Field Validation of the Carbon Sequestration Potential of Coal Seams

in the Central Appalachian Basin, 2011: Blacksburg, VA.

Virginia Department of Mines, Minerals and Energy (VaDMME). 2002. “CBM-PGP-BD114

Completion Report”, Form DGO-GO-15, 105 p.

Harpalani, S, personal communication, December 14, 2012.

Page 61: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

47

MODEL VERIFICATION OF CARBON DIOXIDE SEQUESTRATION

IN UNMINEABLE COAL SEAMS WITH ENHANCED COALBED

METHANE RECOVERY3

Foteini Vasilikou, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Cigdem Keles, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Zach Agioutantis, Department of Mineral Resources Engineering,

Technical University of Crete

Nino Ripepi, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Michael Karmis, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Abstract

Commercial deployment of Carbon Capture Utilization and Storage (CCUS) requires field

testing of a scale that can stress the geologic reservoirs. One such reservoir of interest consists of

unminable coal seams that exhibit favorable characteristics and depositional environments and

lower pressure and temperature than other, deeper, reservoirs. Such conditions can reduce

compression costs while utilizing the action of adsorption that offers a more effective carbon

dioxide (CO2) bonding than free storage or solution. To ensure the success of such tests, a

3 Model Verification of Carbon Dioxide Sequestration in Unmineable Coal Seams with

Enhanced Coalbed Methane Recovery.

F. Vasilikou, C. Keles, Z. Agioutantis, N. Ripepi, M. Karmis 2013 23rd

World Mining

Congress, Montreal, Canada.

Used with permission of the Canadian Institute of Mining, Metallurgy and Petroleum.

Foteini Vasilikou researched and prepared this manuscript, with Cigdem Keles, Zach

Agioutantis, Nino Ripepi and Michael Karmis providing technical and editorial input.

Page 62: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

48

number of parameters should be accurately determined, such as the seam geometry and

stratigraphy, coal porosity and permeability parameters as well as optimum injection conditions.

In essence the CO2 injection model should be calibrated for the reservoir characteristics using all

available data. History matching calibrations are based on gas and water production from

existing wells prior to injection. This paper will present reservoir models that were developed for

a pilot--scale 907 tonnes (1000 tons) CO2 injection test that was performed in 2009 through one

legacy coalbed methane production well in Russell County, Virginia, USA. The model

incorporates a number of individual coal seams about 0.3 m (1ft) in thickness located at depths

ranging from 300 to 700 m (1,000 to 2,200 feet). Model calibration was performed through

history matching to prior production data, and subsequently the model was utilized to develop

different injection scenarios. The developed model was used after injection to calculate CO2

plume distribution patterns that were monitored at the injection test. The paper will present

model data and assumptions with a special emphasis on the effect of hydraulic fractures and the

skin factor to the coalbed methane (CBM) production model.

Introduction

The mitigation of greenhouse gas emissions and enhanced recovery of coalbed methane are

benefits to sequestering CO2 in coal seams. This is possible because of the affinity of coal to

preferentially adsorb CO2 over methane (Shi and Durucan, 2005). Coalbed methane (CBM) is

the most significant natural gas reserve in the Virginia portion of the Central Appalachian Basin,

USA and currently is economically produced in many fields in the Basin from vertical CBM

wells. The recovery factor for vertical CBM well development is estimated that 55% of the gas

in place would be recovered by primary recovery techniques, 20% of the gas in place is

unrecoverable residual gas, and the remaining 25% can be recovered by implementing CO2-

sequestration operations. Since coal has a greater affinity for CO2 than for methane gas, the

injected CO2 should preferentially be adsorbed on the surface of the coal, thereby releasing

methane gas that would be recovered at offset producing CBM wells. A field verification test

successfully injected 907 tonnes of CO2 into a mature CBM production well at the Russell

County, Virginia test site hosted by CNX Gas. The injection commenced on January 9, 2009 and

was completed on February 10, 2009. The maximum daily injection rate was over 50 tonnes of

CO2 per day, with an average injection rate above 36 tonnes per day.

Page 63: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

49

Field Description

The Central Appalachian Basin is a northeast-to-southwest trending basin encompassing

approximately 25,900 km2 in southwestern Virginia, southern West Virginia and eastern

Kentucky (Conrad, 2006). Production of CBM began in 1988 with the development of the Nora

Field in Dickenson County, Virginia followed by CONSOL Energy / CNX Gas developing the

Oakwood Field in Buchanan County, Virginia. Since that time, over 5,600 CBM wells have

been drilled and brought on-line as producing gas wells in southwest Virginia through 2012

(VaDMME, 2013). As of year-end 2010, the coal seams in the Central Appalachian Basin had

produced over 28 km3 of CBM (VaDMME, 2013). Virginia is the primary producer of CBM in

the basin accounting for over 90% of the production. In 2010, Virginia produced a record 3.4

km3 of CBM which accounted for nearly 80% of the natural gas produced in the Commonwealth

(VaDMME, 2013). The U.S. Department of Energy (2013) stated that Virginia accounted for

5.1% of the CBM production in the U.S. in 2010 and accounts for 10% of the CBM reserves,

nearly 57 km3. The majority of CBM development is in areas where gas contents range between

12.5 – 18.7 cubic meters of gas per tonne of coal (Conrad, 2006), making these seams some of

the gassiest in the country. The CBM productivity of the basin indicates that coal permeability

should allow for carbon dioxide injection and storage (Karmis, 2008).

The coals in the region include those of the Lee Formation and Pocahontas Formation and

are medium to low-volatile bituminous, high rank and high gas content coals. In the area of the

injection, there are multiple thin unmineable coal seams (up to 24 separate coals) with net coal

thickness of up to 11 m. The coals average less than 0.6 m in thickness and are deposited over a

large range in depth, 250 to 700 m.

Based on geologic characterization and production studies, wells in the South Oakwood

CBM Field were deemed appropriate for the injection test. These CBM wells were developed on

242,800 m2 grid spacing from 2002 to 2005. Stimulation occurs through perforations in the well

casing into coals greater than 0.15 m thick with a multiple-stage nitrogen foam hydraulic

fracturing treatment (VaDMME, 2002). The purpose of hydraulic fracturing is to breakdown the

coal and create fractures that allow gas to flow through the coal matrix to the fractures and then

to the wellbore. In order to keep these fractures open for gas to flow to the well bore, a proppant

Page 64: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

50

of coarse sand was injected with the high pressure nitrogen foam. Water is produced through a

73.025 mm string of tubing set below the deepest coal seam, and gas is produced between the

casing and the tubing (annulus). The water is removed from the formation to decrease the

pressure of the bed and to allow for the methane to be desorbed and be produced through the

well (Holman, 1996).

CNX Gas donated an existing CBM well in the South Oakwood Field in Russell County,

Virginia for the carbon dioxide injection field test. The injection well is referenced by its well

name, BD114, or by its designation assigned by the State of Virginia, RU-84. Prior to injection,

BD114 averaged production of 1189 m3 of natural gas per day and 0.26 m

3 of water per day

since 2002 and is a slightly below average gas producer for this gas field.

Field Site Layout

A coal field cleat examination was completed to verify the cleat direction of coals at the

injection site verses known face and butt cleat directions of N18W and N67E respectively, for

the deep mined Pocahontas No. 3 coal seam in Buchanan County. Face and butt cleat planes

were measured at two coal seam outcrop locations near the site and they corresponded to within

7 degrees of the known values. A Rose Diagram (Figure 22) was developed to graphically

display the cleat directions from the injection well (VCCER, 2011).

Prior to injecting carbon dioxide, two monitor wells were drilled in close proximity to the

injection well (BD-114), one 41.2 m away (M1) and the other 87.5 m away (M2). The two

monitor wells were drilled at roughly 90 degree offsets from the injection well through the

deepest coal seam that had been perforated in the injection well and exposed each coal seam

being injected into using a packer and tubing. These monitor wells were used to monitor the

pressure of plume, as well as composition of gas at each well. This study affirms that the

monitoring wells for the test site are arranged in both the hydraulic fracture (M1) and face cleat

(M2) directions (Figure 22).

CO2 Injection

The field test successfully injected 907 tonnes of CO2 at the Russell County, Virginia test

site hosted by CNX Gas. The injection commenced on January 9, 2009 and was completed on

February 10, 2009. The data was gathered on a minute basis throughout the one-month injection

Page 65: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

51

and then averaged on an hourly basis to smooth out the curves. Throughout the injection, the

temperature of CO2 injection was maintained close to 100 degrees Fahrenheit, the pressure was

held below 6895 kPa (1,000 psia) and the flow rate varied according to CO2 delivery availability

and operating below the maximum pressure. The maximum daily injection rate was over 50

tonnes of CO2 per day, with an average injection rate above 36 tonnes per day. During the last

three days of injection, at close to maximum pressure, the injection rate declined to a low of 15.4

tonnes per day. The decrease in the injection rate could be attributed to either the pressure of the

reservoir pushing back or swelling of the coals due to adsorption of CO2 at the higher pressures.

Results from the injection provided essential data for establishing the conditions under which

CO2 can be injected into underlying coal seams, ultimately to establish a reasonable estimate of

the volume of CO2 that can be sequestered in Central Appalachian coal seams through predictive

reservoir modeling.

Figure 22 - (A) Rose Diagram Showing Face and Butt Cleat Orientation (B) Rotation by

35o degrees clockwise

Injection Logging

While the CO2 was being injected at a defined temperature, pressure, and rate, a spinner

survey was run downhole to establish the quantity of CO2 being injected into each coal seam. As

part of the spinner survey, temperature and pressure were logged via a wireline downhole to help

profile the injection operations. This survey encountered problems in that the spinner ran into

liquid CO2 at 506 m deep in the wellbore. The temperature log shows the sudden decrease in

0

90

180

3

2

1

2

11

-114

- Average Butt Cleat Orientation: 74o

Average Face Cleat Orientation: 167o

Direction of Max Horizontal Stress: 55o

M1

M2

90o

180o

270o

360o

BD-114/RU-84

0

90

180

3

2

1

2

11

-114 -

Butt CleatsFace Cleats

Direction of Max Horizontal Stress

M1

M2

90 o

180 o

270 o

360 o

BD-114/RU-84

J Direction

I Direction

Page 66: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

52

temperature which happened when the CO2 changed phases from gas to liquid. The results of the

temperature log show significant changes in temperature occurring at each open perforation

where CO2 was injected into coal seams. The change in temperature was plotted and shows that

the greatest change occurred at the shallowest seams that were receiving CO2. The change in

temperature most likely can be correlated with flow rate (VCCER, 2012).

Monitoring Well Results

As the CO2 injection commenced, it was obvious that there was a direct connection through

existing hydraulic fractures to the closest monitoring well, M1. Within 30 minutes of starting the

injection operation, the pressure in M1 unexpectedly raised rapidly to 3,447 kPa (500 psia) and

CO2 content increased to greater than 95%. The profile of the pressure in M1 followed closely

with the pressure at the injection well, lagging about 690 kPa (100 psia) through the one month

injection. At the end of the injection, the pressure in M2 mirrored M1 and the injection well. It is

unknown if the rise in pressure in M2 was as quick as M1, but the results indicate that the M2 is

also interconnected in a fracture network with the injection well, BD114 as the CO2 content in

M2 also reached greater than 95%.

Reservoir Modeling

The CBM production data between 2002 and 2009 was used to develop a reservoir model of

the area. The original model was developed utilizing Advanced Resources’ COMET3 reservoir

simulator for CBM (Ripepi et al., 2009), but in 2013 a new model was developed using

Computer Modeling Group Ltd.’s GEM software, a compositional and unconventional reservoir

simulator (Vasilikou et al., 2013), so that the effect of fractures could be incorporated. The 2013

study utilized a different geometry and reservoir parameters resulting in a better history match

than before, while obtaining insight regarding the specific reservoir characteristics and behavior.

This paper complements existing work by introducing the effect of hydraulic fractures to the

CBM production model by employing two methods: using the “Skin Factor” (SF) approach and

using the “Hydraulic Fracture” (HF) approach. Subsequently the central production well is

converted to an injector and CO2 is injected into the reservoir for a month according to the

injection history presented in the previous section.

Page 67: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

53

Theoretical Considerations for the Skin Factor and the Hydraulic Fracture Approach

The permeability near the wellbore is different from the permeability deeper in the

formation. It is either reduced as affected by drilling activities or it is enhanced due to well

stimulation. This reduced or enhanced permeability zone, termed as “skin” zone, causes an

additional pressure change when compared to the pressure drop due to the original permeability

of the formation. According to Hawkins (1956), the additional pressure change across this zone

can be approximated by Darcy’s equation by introducing a skin factor term, defined as follows:

[

] ( ) (26)

When the permeability around the wellbore is higher than the formation permeability, i.e.,

due to well stimulation, , the first term if eq (26) is negative. Since the ( ) term

is always positive, in cases of stimulated wells the skin factor is always negative. In GEM, the

skin factor is incorporated in the molar flow rate equation of a particular layer into the well. The

extent of the effective radius ( ) for each well can be calculated using the following equation

(Peaceman, 1983). For the case of the SF Model, a skin factor of -4 was used for all the wells in

the modeling area.

√ √ (27)

Darcy’s equation is employed in order to simulate laminar flow of fluid through porous

media. However, as it has been reported in literature, for hydraulically fractured wells, gas flow

velocities near the wellbore and the pressure drop, are not proportionally increased. For this

reason in order to account for both laminar and turbulent phenomena the following equation

suggested by Forchheimer (1914) is used:

(28)

where

is the pressure gradient, μ the viscosity, k the formation’s permeability, ρ the fluid’s

density, and β the non-Darcy flow coefficient, a factor characteristic of the porous medium. The

β factor is dependent on the gas saturation in the hydraulic fracture, the relative permeability of

gas flow and the fracture porosity. Several authors have published different correlations of the β

factor. In GEM, the following models can be applied in order to determine the β factor: The

Geertsma, Frederick and Graves I or II or a general non-Darcy correlation model.

Page 68: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

54

Another important aspect of accurately simulating a non-Darcy flow response in the

hydraulic fractures is the use of a pseudo fracture. More specifically, in order to model an actual

fracture width of 0.0058 meters, as in the case of the HF model, very high fracture permeability

(i.e. 60,000md) is assigned to that narrow fracture. This is subsequently mapped to the modeled

fracture width, i.e. 0.4m and converted to effective permeability by observing the equation k x

width = constant.

The Importance of Relative Permeability

The relative permeability (RP) is defined as the ratio of the effective permeability (EP) to a

given fluid at a definite saturation to the permeability at 100% saturation (k). Since k is a

constant for a given porous material, the RP varies with the fluid saturation in the same fashion

as does the EP. The RP to a fluid will vary from a value of zero at some low saturation of that

fluid to a value of 1.0 at 100% saturation of that fluid. In this case study, only two phase flow in

the cleats needs to be considered: flow of water and flow of gas (CH4). In the current case study,

the initial water saturation in the cleats is assumed high and numerically is assigned between 0.9

and 1.0. The RP curves are then optimized to work with the permeability of the strata to yield the

appropriate water and gas rates. Figure 23 presents a typical example of the RP curves for two of

the wells in the modelled area. In addition, Figure 24 compares the RP curves used in the SF and

HF approaches.

Figure 23 - Typical Relative Permeability Curves for Two Wells

Page 69: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

55

Figure 24 - Relative Permeability Curves for the HF and SF Models

History Match

A dual porosity, single permeability model was setup, since coal matrix permeability is

considered extremely low. The 24 coal seams encountered at the test site were grouped in four

flat zones set at the average elevation of each group (Vasilikou et al., 2013). The generated grid

encompasses all production wells allowing for a distance of 150-200 m between each well and

the grid boundary. The grid size was set to 25mx25m, with the grid extending to 84 blocks in the

I (x) direction and 56 blocks in the J (y) direction.

The orientation of the hydraulic fractures was critical in determining grid generation. As the

GEM model can only allow for hydraulic fractures along either the I (x) or J (y) direction, and

hydraulic fractures were developed along the maximum horizontal stress in the area, the grid was

rotated by 35 degrees clockwise to orient the maximum horizontal stress in the I direction

(Vasilikou et al., 2013). In addition, it was assumed that the permeability in the I direction is

equal to the butt cleat permeability and that the permeability in the J direction is equal to the face

cleat permeability. Furthermore, the modeled group of eight wells was considered independent of

other wells surrounding this grid. The resulting grid for all layers including the eight wells is

shown in 3D view in Figure 25. The face cleat permeability distribution for all layers varies

between 8 and 12 md variation. Similar results are obtained for the SF and HF models, where the

face cleat permeability values are in the same range as before, but different since these reservoir

models exhibit different flow characteristics.

Page 70: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

56

Figure 25 - Face Cleat Permeability Variation for the

Base Model

Figure 26 - Typical HF

Discretization for All

Wells

Figure 26 presents a detail of the model at the central production / injection well where the

hydraulic fracture is shown as an area with finer grid discretization. The effective permeability

for that area is in the order of 57,000 mD. It should be noted that these are optimum permeability

values as determined by optimization of the reservoir properties. The RP curves varied per well

for the base model, while they were kept constant for the SF and HF models since the later allow

for higher permeability values in the near well bore areas. Adsorption isotherms for CO2 and

CH4 were developed by a commercial laboratory, Pine Crest Technology, on three coal seams

that are representative of coals developed for CBM in the South Oakwood CBM field.

In all above runs, the primary operating constraint was to meet the gas rate, while the

secondary operating constraint was that the bottom hole pressure should not fall below a

minimum value of 200 kPa. The objective function for the optimization was to minimize the

error for the cumulative water that is recorded per well. It should be noted that the water output

logged contains water introduced into the well and the formation during the hydraulic fracturing

procedure.

Results and Discussion

Figure 27 compares the cumulative water produced at well BC-115 (RU-123) over a six year

period. The dots represent the measured values, while the dashed lines represent the calculated

values by the three models discussed previously. In this case all models over-predict the water

Page 71: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

57

produced. Although only one such diagram is presented in this paper, such charts can be

composed for all wells. For other wells, the calculated cumulative water does not follow the

same pattern and some models under predict while others over predict the measured values. It

should be emphasized, that a complex problem like water and gas production from multiple

wells may have more than one solution depending on the parameters selected. In fact, the

solution space can be considered practically infinite. In addition, the HF model requires more

computer resources to complete than the other models, while it is also the most sensitive of the

three when input parameters vary. In all cases the models matched the historic data for the gas

rate very well.

Figure 27 - Cumulative Water Production vs. Measured Data for Well BC-115 (RU-123)

for All Models

Following the production simulation, and the respective history match, CO2 was injected

into well BD 114 (RU 84) for a period of about one month. The actual injection data were

available on a daily basis, while the production data were available on a monthly basis. Figure 28

presents the injected gas rate (surface conditions, m3/day) and the injected cumulative gas

(surface conditions, m3) for the injection period. Following injection, the injector well was shut

in for a few months.

Page 72: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

58

Figure 28 - Gas Rate and Cumulative Gas during CO2 Injection into Well BD 114 (RU 84)

Figure 29 shows the spatial distribution of the CO2 plume around the injector well

immediately after injection has been completed for the HF model. As expected, the plume travels

along the high EP zones on either side of the well. In addition the CO2 plume decreases in

magnitude as depth increases. Results indicate that the topmost injection zone received more

CO2 than the injection zone at the bottom. In addition, the simulated CO2 plume reaches the

location of monitoring well M1 which was drilled in the path of the hydraulic fractures (see

Figure 22 B). The SF and base models show a uniformly distributed CO2 around the well bore.

Figure 29 - CO2 Plume around Injection Well Immediately After Injection Has Been

Completed In the HF Model

Page 73: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

59

Conclusions

Due to the complexity, non-linearity and parameter variability of CBM reservoir modeling

multiple solutions may be obtained that satisfy input parameters, boundary conditions, solution

constraints and the objective function. A successful model should predict future production after

being calibrated using historic data. In this work, two advanced CBM production models (the

Skin Factor model and the Hydraulic Fracture model) were developed by incorporating the

effects of permeability changes due to hydraulic fracturing. The models are compared to each

other and also compared to a base model that does not incorporate any effective permeability

changes due to fracturing.

The advantage of the SF model is that it can run in much less time than the HF model. The

skin factor approach however induces a symmetric effective permeability change around each

well since the skin factor is not directional. This is also evident in the plume representations for

the SF models. In contrast the HF models correctly account for preferential effective

permeability changes around the fractured area, but take a lot more computer resources to

complete. Although all models converge and meet the specified gas rate and cumulative gas

production values, none of the models provides a unique match to water production data for all

wells.

These preliminary results verify that the injection process characterized by sorption and the

process of flowback followed by CH4 and/or CO2 desorption is mainly driven by the effective

permeability of the formation around each well. The skin factor approach may be applicable for

large scale simulations where well interference is not an issue, while the hydraulic fracture

simulation should be used when more accuracy in the formation and migration of injected and/or

produced gases is needed.

Acknowledgments

Financial assistance for this work was provided by the U.S. Department of Energy through

the National Energy Technology Laboratory’s Program under Contract No. DE-FC26-

04NT42590 and DE-FE0006827.

Page 74: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

60

References

GEM (Version 2011.12) [Computer software]. Calgary, Canada: Computer Modelling Group Ltd

(CMG).

Conrad, J.M., Miller, M.J., Phillips, J. & Ripepi, N. (2006). Characterization of Central

Appalachian Basin CBM development: Potential for carbon sequestration and enhanced

CBM recovery, 2006 International Coalbed Methane Symposium, Preprint 0625,

Tuscaloosa, AL.

Conrad, J.M., Miller, M.J. & Ripepi N. (2007). Potential for carbon sequestration in the Central

Appalachian Basin. Air and Waste Management Association Conference, 2007, Paper # 187.

Forchheimer, P. (1914). Hydraulik, Teubner, Leipzig and Berlin, 116-118.

Hawkins, M. (1956). A note on the skin effect. Trans. AIME, 207, 356-357.

Holman, T. (1996). Analysis and optimization of coalbed methane gas well, production (master’s

thesis). Retrieved from Virginia Polytechnic Institute and State University, Blacksburg, VA.

Karmis, M., Ripepi, N., Miskovic, I., Conrad, J.M., Miller, M.J., & Shea, C. (2008). CO2

Sequestration in unminable coal seams: Characterization, modeling, assessment and testing

of sinks in Central Appalachia. Twenty-Fifth Annual International Pittsburgh Coal

Conference, Pittsburgh, PA, USA, 2008. CD-ROM of Proceedings, Paper # 29-3.

Peaceman, D.W., (1983). Interpretation of well-block pressures in numerical reservoir simulation

with non-square grid blocks and anisotropic permeability. Society of Petroleum Engineers

Journal, 23, 531-543.

Ripepi, N., Karmis, M., Miskovic, I., Shea C., & Conrad, J.M. (2009). Results from the Central

Appalachian Basin field verification test in coal seams. 26th Annual International Pittsburgh

Coal Conference, Pittsburgh, PA, USA, 2009. CD-ROM of Proceedings, Paper # 33-4.

Shi, J.Q. & Durucan, S. (2005). CO2 Storage in deep unminable coal seams. Oil & Gas Science

and Technology, IFP, 60(3), 547-558.

United States Department of Energy (2013). Coalbed methane proved reserves and production,

Energy Information Administration, accessed February 14: Retrieved from U.S. Energy

Information Administration website:

http://www.eia.gov/dnav/ng/NG_ENR_COALBED_DCU_NUS_A.htm

Page 75: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

61

Vasilikou, F. Keles, C., Agioutantis, Z., Ripepi N. and Karmis, M. (2013). Experiences in

reservoir model calibration for coalbed methane production in deep coal seams in Russell

County, Virginia, Proceedings, Symposium on Environmental Considerations in Energy

Production, April 14-18, 2013, Charleston, West Virginia (in press).

Virginia Center for Coal and Energy Research (VCCER) (2011). Final technical report:

Characterization and Field Validation of the Carbon Sequestration Potential of Coal Seams

in the Central Appalachian Basin, 2011: Blacksburg, VA.

Virginia Department of Mines, Minerals and Energy (VaDMME) (2002). CBM-PGP-BD114

Completion Report, Form DGO-GO-15, 105 p.

Virginia Department of Mines, Minerals and Energy (VaDMME) (2013), accessed February

2013: Retrieved from Virginia Department of Mines, Minerals and Energy website:

http://www.dmme.virginia.gov/DGO/Production/2010-County.pdf

Page 76: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

62

RESERVOIR SIMULATIONS FOR COAL BED METHANE (CH4)

PRODUCTION AND CARBON DIOXIDE (CO2) INJECTION IN DEEP

COAL SEAMS IN BUCHANAN COUNTY, VIRGINIA4

Foteini Vasilikou, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Zach Agioutantis, Department of Mineral Resources Engineering,

Technical University of Crete

Nino Ripepi, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Michael Karmis, Virginia Center for Coal and Energy Research, Virginia Tech,

Blacksburg, VA

Reservoir Model Calibration for Coal Bed Methane Production from Deep Coal

Seams in Buchanan County, Virginia

Abstract

The potential of commercial deployment of CO2 sequestration in unmineable coal seams, as

a measure to mitigate the greenhouse gas effect, is being tested through a series of small- to

medium-scale injection projects in the Appalachian Basin. Indispensable tools for these projects

are reservoir simulations in which the involved processes, both prior to and post injection field

tests are modeled to enhance understanding and to be used as a decision tool. Due to the large

number of modeling input parameters and the high uncertainty in their values determination, for

development of representative models initial model sensitivity analysis of the key input

parameters is required. In this paper, preliminary single-well reservoir parametric simulations are

conducted to identify how the variation of selected reservoir parameters and properties affect

model results. More specifically, this sensitivity analysis focused on three areas: a) the effect of

4 This paper is intended for publication.

Foteini Vasilikou collected the data, researched, and wrote this manuscript with technical

input from Zach Agioutantis, Nino Ripepi and Michael Karmis.

Page 77: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

63

selected initial conditions (pressure gradient, adsorption isotherm, and porosity), b) the influence

of the production mechanism (permeability, relative permeability, and compressibility), and c)

the contribution of the well characteristics (hydraulically stimulated well, skin factor). This

modeling work forms the base for calibration of reservoir models for simulating CO2 injection

and subsequent prediction of post-injection behavior for an upcoming injection test in Buchanan

County, VA.

Introduction

The Coal Seam Group, led by researchers at the Virginia Center for Coal and Energy

Research (VCCER), part of the Southeast Regional Carbon Sequestration Partnership

(SECARB), completed a small-scale validation field test in Russell County, Virginia, during

2009, where 1,000 tons of CO2 were injected into one vertical coalbed methane well over a

period of one month (Ripepi et al., 2009). Prior to conducting the injection test, preliminary

reservoir simulation models were developed to predict the extent of the CO2 plumes and estimate

injection pressures (VCCER, 2011). After the injection test was completed the original reservoir

models were updated to account for post injection data and were refined based on the knowledge

gained (Vasilikou et al., 2013).

In 2014, the VCCER will be performing a larger scale validation test in Buchanan County,

Virginia, approximately seven miles to the north west of the Russell County site. Twenty

thousand (20,000) tons of CO2 will be injected into three vertical coalbed methane wells over a

period of one year. The enhanced understanding of the production and injection mechanisms into

stacked geologic systems gained from the small injection test in Russell County, along with

updated reservoir software capabilities, will be used to develop improved simulation predictions

for the CO2 injection test in Buchanan County.

Geology

CNX Gas Corporation, a subsidiary of CONSOL Energy Inc., has donated three vertical

coalbed methane production (CBM) wells in Buchanan County, Virginia within the Central

Appalachian Basin region. These three production wells will be shut in and converted into

injection wells a month prior to the field injection test start date. As recorded by the State of

Virginia the candidate injection wells, BU 1923, BU 3337, BU 1998 or alternatively as

Page 78: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

64

referenced by their company name, DD7, DD7A, DD8, are at an elevation of 2286.6 feet, 1932.2

feet, and 1876.2 feet, respectively (DMME, 2011).

The Central Appalachian Basin is a northeast-to-southwest trending basin encompassing

approximately 10,000 square miles in southwestern Virginia, southern West Virginia and eastern

Kentucky (Conrad et al., 2006). Production of CBM began in 1988 with the development of the

Nora Field in Dickenson County, Virginia, followed by CONSOL Energy developing the

Oakwood Field in Buchanan County, Virginia. The coals in the region near the injection site

include those of the Pocahontas Formation and Lee Formation, which directly overlie the late

Mississippian Bluestone Formation. Coal seams of the Pocahontas and Lee Formation are

medium to low-volatile bituminous, high rank and high gas content coals that include the

Pocahontas No. 1 through Pocahontas No. 11 seams (Pocahontas Formation and the Greasy

Creek, Middle and Lower Seaboard and Upper Horsepen coals) (Vasilikou et al, 2013).

Cardno MM&A, a research partner on this project, created a database for the coal seams in

the study area by integrating data from donated sources with geophysical well log data from the

Virginia Division of Gas and Oil and the West Virginia Geological and Economic Survey

(VCCER, 2014). This database was used to create elevation structure maps and net thickness

isopachs for both the Lee and Pocahontas Formation coals in southwestern Virginia and southern

West Virginia, encompassing the Central Appalachian Basin. This assessment accounts for the

coal seams that were stimulated for coalbed methane development.

Natural occurring fracture networks in coalbeds and the face and butt cleat planes were

measured at two coal seam outcrop locations near the site and they corresponded to within 7

degrees of the published face and butt cleat directions of N18W and N167E respectively, for the

deep mined Pocahontas No. 3 coal seam in Buchanan County (McCulloch, 1974).

The primary confining units in the study area include multiple layers of low permeability

shale and siltstone beds. Permeability for the shale and siltstone units is expected to range from

0.001 to 0.1 mD, with low porosity (Grimm, 2010). The Lee and Pocahontas Formation

sandstones are known to have low permeability and porosity values, and do not produce natural

gas in this area (VCCER, 2011). Based on well log interpretations, the sandstone units are

expected to have porosity values that range from 1.0 to 3.0 percent and permeability values

ranging from 0.1 to 1.0 mD. All lithologies will encounter some natural fractures; however, these

fractures are likely to be cemented with quartz and calcite and are not expected to provide

Page 79: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

65

permeability pathways based on core analysis, well logging and field testing (Karmis et al.,

2008).

Well Stimulation

All wells within the study area have been perforated and hydraulically fractured using a

nitrogen foam hydraulic fracturing treatment. The wells selected for injection, DD7 (BU 1923),

DD7A (BU 3337), DD8 (BU 1998) were stimulated in November 2000, March 2007 and May

2001, respectively. DD7 was completed in four zones ranging in depth over 1216-1385 feet,

1531-1570 feet, 1674-1866 feet and 1997-2133 feet. DD7A was stimulated in three stages at

856-975.5 feet, 1203-1420 feet and 1599-1803.5 feet. DD8 was completed in three zones ranging

in depth over 871-1205 feet, 1311-1677.5 feet and 1710.5-1802 feet (DMME, 2011).

In each well, the main hydraulic fractures at each perforation are oriented at approximately

N57E (VCCER, 2011). Although in some cases fracturing may extend up to 700-800 feet away

from the borehole, in this analysis a more conservative approach was used where the fracture

half-length is assumed to be 350 feet. In addition, after personal communication with the

operator of the wells, it was suggested that only 40 percent of the coal seams per stage were

successfully completed (CNX, 2013).

Production History

The three injectors, DD7, DD7A, and DD8, were brought on-line as coalbed methane

production wells in 2000, 2007 and 2001, respectively. To date, DD7, DD7A, and DD8 have

averaged a production of 82.52 Mcf, 41.25 Mcf and 47.88 Mcf of natural gas per day and 1.42

barrels, 1.11 barrels and 1.74 barrels of water per day (DMME, 2013). The cumulative gas

production of these three wells to date is estimated to account for 58 percent, 56 percent and 33

percent, respectively, of their expected ultimate gas recovery (VCCER, 2014).

The gas production data are measured at each well, whereas the water production data are

estimated by dividing the total production of a group of wells by the total time the water pump

on each well was operating. For this reason, there is a higher degree of confidence in the

accuracy of the gas production data versus the reported water production of the wells.

Page 80: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

66

Figure 30 shows historic gas production rates for the three selected injection wells, DD7,

DD7A and DD8. Daily water production for the three injectors is shown in Figure 31.

Figure 30 - Daily Gas Production Rate of the Injection Wells

Page 81: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

67

Figure 31 - Daily Water Production Rate of the Injection Wells

Description of Reservoir Model

Reservoir simulations prior to commencing with the CO2 injection field test were performed

to estimate the extent of the CO2 plumes, the storage capacity of unmineable coal seams and the

potential of enhanced gas recovery of the field. The reservoir simulator employed in this study is

the compositional and unconventional simulator GEM by the Computer Modeling Group Ltd.

This paper focuses on studying the interaction of key properties which affect the initial gas

and water volumetrics in coalbeds and the production mechanism from multiple thin coal seams.

In addition, well characteristics relevant to how the well was drilled and perforated, such as the

frequency of hydraulic fractures, the fracture lengths and widths, are examined with respect to

ultimate well productivity. Due to the nature of the multi-seam configuration of the study area

and the degree of interference between input parameters, this paper focuses on part of the study

area that includes a single well, DD7, in order to investigate commingled primary production for

different scenarios. Simulation results of the whole study area with CO2 injection in three of the

wells are detailed below.

Page 82: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

68

The geologic structure employed for the single well simulations in this study is an 18-layer

model (Figure 32A) proposed by Cardno MM&A (2013), a project subcontractor. Table 6 shows

the top elevations and thicknesses for each simulated coal seam. Numbering of the layers starts

with the shallowest seams and increases towards the deepest seams. The drainage area for the

single well simulations is approximately 42 acres.

In addition, based on a slightly different interpretation of the aforementioned geologic

structure proposed by Cardno MM&A, a second structure where 5 zones of coal seams are

assumed is also investigated in order to examine sensitivity of model response to variation of

number of layers in the model (Figure 32B).

Figure 32 - (A) 18-Layer Model, (B) 5-Zone Model

18- Layer Models

The key parameters investigated in this analysis are: (i) pore pressure gradient, (ii) coal and

cleat porosity, (iii) Langmuir constants, (iv) water saturation in cleats, (v) cleat permeability, (vi)

relative permeability curves to water and gas, (vii) methane desorption time from the coal matrix,

(viii) skin factor and (ix) hydraulic fractures for well stimulation. The primary objective in each

model was to match the historic gas and water production up to year 2013 for the selected well

(Figure 33).

(A) (B)

Page 83: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

69

Figure 33 - History Matching of Monthly Gas Rate and Cumulative Gas for Well DD7 Up

to Year 2013

The first set of input parameters that were varied in the simulations is as follows: pore

pressure gradient, coal and cleat porosity, Langmuir volume and Langmuir pressure constants,

and the initial water saturation in the cleats, shown in Table 7. The remaining parameters used in

the first set of models (C) are kept constant as shown in Table 6. Cases C1 through C11

investigate the effect of the aforementioned parameters on the initial gas and water in place of

the models. In cases C1, C2 and C3 the variation of the original gas and water in place is

examined in relationship to the variation of the Langmuir volume constant, VL; in case C4, the

water saturation in the cleats is reduced from 100 to 90 percent and all the other parameters are

the same as in case C1; in case C5 the Langmuir pressure constant, PL, is reduced from 0.01 to

0.003 and the rest of the assigned properties are the same as in case C3; in cases C6, C7, C8 all

the parameters are kept the same as in cases C1, C2, C3 respectively, but the pore pressure

gradient assigned in the fractures is increased from 0.315 psi/ft to 0.36 psi/ft; in cases C9, C10,

C11 the coal and cleat porosity are reduced compared to the respective values in cases C1, C2,

C3 from 1 percent (Pashin, 2014) to 0.15 percent (Liu et al., 2012) and from 0.1 to 0.015 percent.

Figure 34 shows the effect of the variation of the Langmuir constants to the amount of adsorbed

Page 84: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

70

gas per coal mass constrained by the minimum and maximum pore pressures at the shallower

and deeper seam, respectively, depending on the pore pressure gradient employed.

Table 6 - Parameters and Material Properties Used in the 18-Layer Reservoir Model

Number of Grid Blocks per Direction I=27,J=27, K=18 Layers

Grid Block Sizes per Direction, ft DI=50, DJ=50, DK= Variable

Coal Layers and Thicknesses, ft (L1) 0.02-1.38, (L2) 0.71-1.49, (L3) 0.52-1.10,

(L4) 1.31-1.69, (L5) 1.31-1.69, (L6) 0.28-0.79,

(L7) 0.53-0.80, (L8) 0.66-1, (L9) 1.93-2.49,

(L10) 0.01-0.46, (L11) 0.61-1.25, (L12) 0.01-0.87,

(L13) 0.71-1.40, (L14) 0.61-1.19, (L15) 1.61-2.19,

(L16) 0.61-1, (L17) 0.02-1.07, (L18) 0.303-0.526

Coal Layer Elevations-Grid Tops, ft (L1) 1,163-1,178, (L2) 1,067-1,086, (L3) 1,026-1,041,

(L4) 955-972, (L5) 909-934, (L6) 759-776,

(L7) 730-750, (L8) 724-743, (L9) 615-633,

(L10) 612-624, (L11) 512-526, (L12) 327-343,

(L13) 295-318, (L14) 243-267, (L15) 193-219,

(L16) 176-195, (L17) 169-182, (L18) 119-144

Pore Pressure Gradient, psi/ft 0.315

Coal (Matrix) Porosity, % 1

Cleat (Fracture) Porosity, % 0.1

Coal (Matrix) Permeability, mD Not in use, single permeability model (GEM, 2003)

Cleat (Fracture) Permeability, mD I: 30, J: 30, K: 3

Relative Permeability Curves

Gas-Water

Relative water permeability 1 and relative gas

permeability 1 ( Figure 35)

Langmuir Adsorption Constants Langmuir Volume Constant, VL= 800 scf/ton

Langmuir Pressure Constant, PL= 333 psi

Fracture Spacing per Direction, ft I: 0.01667, J: 0.01667, K: 0.01667

Well Radius, ft 0.13124

Formation Compressibility

(Matrix & Fracture), 1/psi

2E-08 (at reference pressure 416.15 psi)

(Harpalani, 2012)

Reservoir Temperature, oF 69.8

Water Density, lb/ft3 62.4779 (at reference pressure 416.15 psi)

Coal Density, lb/ft3 82.65(VCCER, 2011)

Desorption Time, days 20

Cleat (Fracture) Water Saturation 1

Well Stimulation Not in use, zero skin factor, no hydraulic fractures

Page 85: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

71

Table 7 - Set of Parameters Varied in Simulations for Cases C1 to C11

Parameters C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11

Pressure

Gradient,

psi/ft 0.315 0.315 0.315 0.315 0.315 0.36 0.36 0.36 0.315 0.315 0.315

Coal (Matrix)

Porosity, % 1 1 1 1 1 1 1 1 0.15 0.15 0.15

Cleat

(Fracture)

Porosity, % 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.015 0.015 0.015

Langmuir

Constant, VL,

scf/ton 500 650 800 500 800 500 650 800 500 650 800

Langmuir

Constant, PL,

1/psi 0.01 0.01 0.01 0.01 0.003 0.01 0.01 0.01 0.01 0.01 0.01

Water

Saturation in

Cleats, % 100 100 100 90 100 100 100 100 100 100 100

Figure 34 - Langmuir Adsorption Isotherms for Langmuir Volume Constants 500, 650, 800

scf/ton and Langmuir Pressure Constants 100 and 333 psi; Maximum and Minimum Pore

Pressures in the Reservoir for Pressure Gradients 0.315 and 0.36 psi/ft

Page 86: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

72

The second input parameter investigated was cleat permeability. The sensitivity of the 18-

layer model to variation of cleat permeability was examined. Three models - P1, P2 and P3 -

with an isotropic cleat permeability in space and in depth of 20 mD, 30 mD and 45 mD (same as

in cases C1 through C11) were studied. The rest of the parameters were kept constant as shown

in Table 6.

Two different relative permeability curves for water and two for gas were considered in this

analysis as shown in Figure 35. All other input parameters for the models were as shown in

Table 6.

Desorption time from the coal matrix to the cleat system was also varied. In the

investigation of cases C1 to C11 desorption time was set to 20 days. Desorption times of 5 and

50 days was also examined. The rest of the parameters were kept constant as shown in Table 6.

Figure 35 - Different Scenarios for the Relative Permeability Curves to Water and Gas

Because it was hydraulically stimulated, to account for enhanced flow properties around

DD7 three scenarios were studied (S1, S2 and S3). In the first scenario (S1) a skin equal to zero

was assigned; in the second (S2) a negative skin of -2 and in the third (S3) a negative skin of -4

Page 87: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

73

was used. The minimum value of skin factor that could be assigned in order to maintain a

positive production index for the well was -4.

Another way to account for hydraulically stimulated wells in reservoir simulations is to

explicitly model hydraulic fractures. This approach is more computationally intensive compared

to assigning a negative skin factor. For the purposes of this investigation four scenarios with

hydraulic fractures were studied. In all four scenarios (H1, H2, H3, H4) only 40 percent of the

seams per stage, as suggested by the operator of the wells, were assumed to be successfully

fractured. For scenarios H1, H2 and H3 the thickest seams per stage were selected to be fractured

(Pashin, 2014). In scenario H4 the seams per stage selected to be hydraulically fractured were the

ones that initially had a higher amount of gas in place. The properties of the hydraulic fractures

assumed in the modeling work and the layer numbers selected to be fractured per scenario are

shown in Table 8.

Table 8 - Input Parameters for Modeling Hydraulic Fractures in Scenarios H1, H2, H3, H4

H1 H2 H3 H4

Layer Number

with HF

4, 6, 8, 12, 14,

16, 18

4, 6, 8, 12, 14,

16, 18

4, 6, 8, 12, 14,

16, 18

4, 5, 9, 13, 15,

16, 18

Primary Fracture

Width (ft)

0.00833

(GEM, 2003)

0.01042

(EPA, 2004)

0.12500

(EPA, 2004)

0.00833

(GEM, 2003)

Primary Fracture

Permeability (mD)

6,000

(GEM, 2003)

6,000

(GEM, 2003)

6,000

(GEM, 2003)

6,000

(GEM, 2003)

Fracture Half

Length (ft) 300 300 300 300

Grid Refinement I: 9, J: 9, K: 1 I: 9, J: 9, K: 1 I: 9, J: 9, K: 1 I: 9, J: 9, K: 1

Effective Fracture

Permeability (mD) 55 61 405 55

5-Zone Model

Finally, due to restricted computational capabilities with respect to model size, it is common

practice in reservoir modeling when simulating complex geologic structures to aggregate

Page 88: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

74

geologic layers/units into zones in order to reduce the number of elements in the simulation. In

this study a 5-zone geologic structure, as previously mentioned, for the investigated area

(VCCER, 2013) was compared to the 18-layer model employed in the modeling work. Table 9

shows the input parameters for the 5-zone geologic structure.

Table 9 - Parameters and Material Properties Used in the 5-Zone Reservoir Model

Number of Grid Blocks per Direction I=27, J=27, K=5 Layers

Grid Block Sizes per Direction, ft DI=50, DJ=50, DK= Variable

Coal Layers and Thicknesses, ft (L1) 0.01-2.38, (L2) 2.6-6.3, (L3) 3.9-6.7,

(L4) 0.01-1.84, (L5) 5.3-8

Coal Layers Elevations-Grid Tops, ft (L1) 992.8-1011.9, (L2) 960-979.1,

(L3) 615.9-636.7 (L4) 406.4-429.3, (L5) 198.1-218.3

Pore Pressure Gradient, psi/ft 0.315

Coal (Matrix) Porosity, % 1

Cleat (Fracture) Porosity, % 0.1

Coal (Matrix) Permeability, mD Not in use, single permeability model (GEM, 2003)

Cleat (Fracture) Permeability, mD I: 30, J: 30, K: 3

Relative Permeability Curves

Gas-Water

Relative water permeability 1 and relative gas

permeability 1 ( Figure 35)

Langmuir Adsorption Constants Langmuir Volume Constant, VL= 800 scf/ton

Langmuir Pressure Constant, PL= 333 psi

Fracture Spacing per Direction, ft I: 0.01667, J: 0.01667, K: 0.01667

Well Radius, ft 0.13124

Formation Compressibility

(Matrix & Fracture), 1/psi

2E-08 (at reference pressure 416.15 psi)

(Harpalani, 2012)

Reservoir Temperature, oF 69.8

Water Density, lb/ft3 62.4779 (at reference pressure 416.15 psi)

Coal Density, lb/ft3 82.65(VCCER, 2011)

Desorption Time, days 20

Cleat (Fracture) Water Saturation 1

Well Stimulation Not in use, zero skin factor, no hydraulic fractures

Results and Discussion

Initial Volumetrics

The results of the parametric analysis for the (C) models are shown in Figure 36 and Figure

37. For cases C1, C2 and C3, where the Langmuir volume constant was varied as shown in Table

7, the original gas in place in case C1 increased by 29.92 and 42.87 percent compared to cases

C2 and C3, respectively. The original water in the system, shown in Figure 37, for cases C1, C2

Page 89: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

75

and C3 was not affected by the variation of the Langmuir constant volume and it was the same

for all cases.

In case C4 the initial water saturation in the cleat system was reduced from 100 to 90

percent. A negligible increase in the original gas in place was noted of approximately 0.42

percent in case C4 compared to case C1. The small increase was due to the increase of free gas in

the cleat system. The original water in place was affected the most by the reduction in the initial

water cleat saturation. In case C4 the original water in place reduced by 10 percent compared to

case C1 (Figure 37).

In cases C5 and C3 all input parameters are the same with the exception of the Langmuir

pressure constant, which in case C5 is reduced from 0.01 1/psi to 0.003 1/psi. The change in the

Langmuir pressure constant had an effect on the curvature of the adsorption isotherm as shown

in Figure 34 and, for the given pore pressure range in the reservoir, it resulted in a reduction of

the original gas in place in the system of approximately 28.22 percent (Figure 36). As expected,

the original water in place of the system was not affected by the reduction in the Langmuir

pressure constant.

In cases C6, C7, C8 a higher pore pressure gradient of 0.36 psi/ft compared to 0.315 psi/ft of

cases C1, C2, C3 is assigned in the models and all other input parameters are assumed to be the

same. The increased pore pressure gradient used in cases C6, C7 and C8 resulted in an increase

by 2.22 percent of the original gas in place compared to cases C1, C2 and C3. There was no

change in the original water in place because of the increase in pore pressure.

In cases C9, C10 and C11, both the coal and cleat porosity of the models were reduced to

0.15 percent and to 0.015 percent respectively, compared to cases C1, C2, C3. Reduction of the

coal porosity results in a negligible increase of the original gas in place (Figure 36). Reduction in

the cleat porosity has a significant effect to the original water in place (Figure 37). It results in a

decrease of 85 percent of the overall water in the system from cases C1, C2, and C3 to C9, C10

and C11.

In all cases aforementioned more than 99 percent of the original gas in place is stored in an

adsorbed state on the coal matrix. Only in case C4, where an initial cleat water saturation less

than 100 percent was assigned, there was more than 1 percent free gas in the cleat system. In the

models it is assumed that no methane is dissolved in the water.

Page 90: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

76

Figure 36 - Original Gas in Place in the Model and Initially Adsorbed Gas on the Coal

Matrix for All Cases (C)

Figure 37 - Original Water in Place for All Cases (C)

Figure 38 presents the coal matrix volume profile development with depth at the different

seams elevations for the 18- layer geologic structure for cases C3 and Case 11. Reduction in both

the coal and cleat porosity in case C11 compared to case C3 results only in 0.08 percent increase

of the coal (matrix) volume per seam in case C11 (Figure 38).

Page 91: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

77

Figure 38 - Coal Matrix Volume per Layer Starting from the Shallowest (Layer 1) to the

Deepest (Layer 18) Seam for Cases C3 and C11

Figure 39 - Minimum and Maximum Pore Pressures per Layer for Two Different Pressure

Gradients Used In C Cases

Page 92: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

78

For the coal seams in the study area in Buchanan County, VA, the pore pressure gradient is

assumed to range between 0.315 and 0.36 psi/ft (VCCER, 2013). By employing the 18-layer

geologic structure proposed, the pressure profile per depth per coal seam for the two different

pore pressure gradients of 0.315 psi/ft and 0.36 psi/ft is developed as shown in Figure 39. For the

0.315 psi/ft pressure gradient assumed, pore pressures in the model range between 239- 692 psi.

For the 0.36 psi/ft gradient, pore pressures in the 18-layer model vary between 274-791 psi.

Figure 40 - Initially Adsorbed Gas per Layer for Cases C3, C5 and C8

The initial coal matrix volume per seam (Figure 38) with specific adsorption properties

assigned and the initial pore pressures per layer (Figure 39) shape the initially adsorbed gas

profile in depth for the 18-layer model. Figure 40 presents the initially adsorbed gas per coal

seam for cases C3, C5 and C8. Initially adsorbed gas for case C8 is higher than in case C5 and

significantly higher compared to case C3. In case C8, the lower Langmuir pressure constant used

shifts the adsorption isotherm towards higher gas contents per coal mass and, with the pore

pressures range shifted to the right as shown in Figure 34, overall higher originally adsorbed gas

in the model compared to cases C5 and C3 (Figure 40) was expected. Layers 9 and 15 have the

highest initially adsorbed gas for all cases. It must be noted that coal matrix volume is

Page 93: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

79

significantly higher for layers 9 and 15 compared to the other layers. For layers 4 and 5, while

coal matrix volumetrics for layer 4 is greater than layer 5, layer 5 is a deeper seam with a higher

pore pressure and thus has a greater initial amount of adsorbed gas. This is also the case for

layers 2 and 13 with similar coal matrix volumetrics, but due to the significant difference in pore

pressures the deepest seam has higher initial adsorbed gas content.

Production Mechanism

In the analysis of the (C) models, variation of the input parameters under consideration

affected the initial volumetrics of the models. Parameters such as cleat permeability, relative

permeability curves to water and gas, and desorption time have an effect on the production

mechanism of coalbed wells.

Three different scenarios for cleat permeability were examined. In scenarios P1, P2 and P3,

cleat permeability of 20, 30 and 45 mD was respectively assigned, with a primary operational

constraint to meet the historic water rate and with a secondary constraint of minimum bottom

hole pressure. The results showed that the percentage of gas produced over the initial gas in

place for scenarios P1, P2 and P3 was 35.40, 35.40 and 39.84 percent, respectively. There is

essentially no change in the recovery factor between scenarios P1 and P2, whereas an increase of

12.32 percent in gas production is noted for scenario P3. The layers in order of highest to lowest

production were as follows: layer 9, 15, 5, 4, 13, 2, 11, 16, 3, 8, 17, 14, 7, 6, 18, 1, 12, and 10. As

expected, the layers with the highest initial gas content are also the ones producing the most gas.

Two different curves for relative permeability to water and two curves for relative

permeability to gas were studied (Figure 35). In coalbed reservoirs primary dewatering of the

seams is important for gas production. Hence, the magnitude of the relative permeability to water

for high water saturations is important for gas production. When water and gas start to co-exist in

the cleats, the shape of the relative permeability curves has a significant effect on the production

profile. In this analysis, when relative permeability to water 1 was changed to relative

permeability to water 2 (Figure 35), more water was initially produced but gas production did not

start until lower water saturations and ultimately gas deliverability of the system were reduced.

When relative permeability curve to gas was changed from (1) to (2), as shown in Figure 35,

initial water deliverability was unaffected and gas production increased.

Regarding variations in desorption time of 5, 20 and 50 days, the results show that the

percentage of gas recovery over the initial gas in place are 39.94, 39.84 and 39.64 percent,

Page 94: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

80

respectively. It was expected that the lower and hence faster desorption time used in the

modeling work would yield higher gas production, but it is not significant.

From the investigation of the (S) models where a skin factor was used to account for

enhanced flow around the well, it was shown that recovery factors of the original gas in place for

scenarios S1 (zero skin), S2 (negative 2) and S3 (negative 4) were 38.82, 41.59 and 45.77,

respectively. The recovery factors in the (S) models were higher compared to the ones of the (H)

models where hydraulic fractures were explicitly simulated. For the H models, recovery factors

for scenarios H1, H2, H3 and H4 were 38.00, 38.68, 40.50 and 38.38 percent, respectively. It

must be noted that even scenario S1, where a zero skin factor is used, has a higher recovery

factor than most of the H models. This is most likely due to the different flow properties assumed

for the H models versus the S models. For the H models, it must be noted that increase in the

primary fracture width (Table 8) yields an increased effective permeability and ultimately a

higher recovery factor. Also, in scenario H4 where hydraulic fractures per stage were selected

based on higher initial gas content, the recovery factor was slightly higher compared to the

recovery of scenario H1 where stimulated seams were selected based on the thickest seams per

stage criterion.

Figure 41 - Originally Adsorbed Gas per Layer for the 5-Zone Model

Page 95: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

81

Figure 42 - Originally Adsorbed Gas per Layer for the 18-Layer Model

The results regarding the two different geologic structures considered, the 5-zone model and

the 18-layer model, showed that the original gas in place was 6.95e5 and 6.38e5 Mcf,

respectively. The gas recovery factor for the 5-zone model was higher compared to the 18-layer

model, which was expected since the original gas in place was also higher.

Conclusions

This paper discusses in detail a sensitivity analysis with respect to key input modeling

parameters for the injection site in Buchanan County, VA. Through parametric analysis, the

effect of several input parameters is determined. Identifying the contribution of each parameter is

important in deciding the optimal modeling configuration and approach for the full field scale

simulation of the CO2 injection in unmineable coals seams for the Buchanan County site.

Page 96: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

82

CO2 Injection Model for Enhanced Coalbed Methane Recovery in Deep Coal

Seams in Buchanan County, VA

Abstract

This paper presents preliminary reservoir simulations which were conducted to predict the

amount of carbon dioxide that can be stored in multiple thin coal seams following injection, the

extent of CO2 plumes by seam, the potential for CO2 breakthrough at offset wells, and the

amount of CO2 stored if the injection wells are flowed back. Finally, the paper will present an

assessment of the potential of enhanced gas recovery for all wells within the study area.

Development of Reservoir Models

Reservoir simulations prior to the commence of the CO2 injection field test were performed

to estimate the extent of the CO2 plumes and identify suitable locations for monitoring wells, the

storage capacity of unmineable coal seams and the potential of enhanced gas recovery of the

field. The simulation covers a drainage area of approximately 1552 acres and incorporates the

three selected injection wells and seventeen offset CBM wells. Table 10 shows the designations

of these 20 wells as well as the date that production started at each well. The date shows that the

wells in the simulated area were brought on-line at different times within a time span of sixteen

years; the first well started production in 1993 and the latest producer started in 2009.

Page 97: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

83

Table 10 - Reference Names and Production Start Dates for the Coalbed Methane Wells

Included in the Simulations

State Well Name Company Well Name Production Start Date

BU 1480 CC6 10/1/2000

BU 3850 CC6A 06/1/2008

BU 1875 CC7 11/1/2000

BU 3335 CC7A 05/1/2007

BU 0377 CC8 02/1/1993

BU 0378 CC9 02/1/1993

BU 3388 CC9B 09/1/2007

BU 1950 DD5 06/1/2001

BU 1743 DD6 12/1/2000

BU 3893 DD6A 08/1/2008

BU 1923 DD7 12/1/2000

BU 3337 DD7A 04/1/2007

BU 1998 DD8 06/1/2001

BU 3929 DD8A 03/1/2009

BU 2092 DD9 11/1/2001

BU 4269 DD9A 09/1/2009

BU 1924 EE6 01/1/2001

BU 1922 EE7 01/1/2001

BU 2072 EE8 09/1/2001

BU 3942 EE9A 03/1/2009

Page 98: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

84

The reservoir simulator employed in this study is the compositional and unconventional

simulator GEM by the Computer Modeling Group Ltd. A dual porosity, single permeability

model is used for modeling the coalbed reservoir, since coals consist of micro and macro pores

and have a negligible matrix permeability compared to their fracture permeability (CMG, 2003).

In the reservoir models, a geologic structure for the coal seams is assumed, based on

integrated data for coalbeds within the study area (VCCER, 2013). Surface elevations in the

study area range from 1685- 2364 feet. The models include 18 layers at elevations ranging from

29 feet to 1201 feet with varying thicknesses between 0.01 to 4.2 feet, as shown in Table 11. It

must be noted that coal seam thicknesses below 0.5 feet are based on an estimation of where the

coal seam will pinch out. Layers in Table 11 are listed from top elevation to bottom elevation.

Page 99: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

85

Table 11 - Layers Number, Coal Seam Names, Thicknesses and Top Elevations for the

Geologic Structure Assumed for the Study Area

Layers Number Coal Seam Coal Thicknesses

(ft)

Coal Elevations

(ft)

1 Greasy Creek 0.03-2.90 1087-1201

2 Seaboard (20) 0.01-1.80 982-1125

3 Lower Seaboard 0.02-1.98 944-1067

4 Upper Horsepen (10) 0.37-2.85 887-992

5 Upper Horsepen (20) 0.03-1.40 829-959

6 Pocahontas No. 11 0.10-1.79 694-816

7 Pocahontas No. 10 (10) 0.01-2.07 672-780

8 Pocahontas No. 10 (20) 0.18-1.54 665-774

9 Pocahontas No. 9 (10) 1.00-3.30 593-633

10 Pocahontas No. 9 (20) 0.10-1.30 591-624

11 Pocahontas No. 7 (10) 0.12-2.47 452-564

12 Pocahontas No. 5 (10) 0.06-3.10 270-372

13 Pocahontas No. 5 (20) 0.20-4.20 234-354

14 Pocahontas No. 4 (20) 0.20-3.70 199-309

15 Pocahontas No. 3 (10) 0.06-3.00 150-266

16 Pocahontas No. 3 (40) 0.10-1.40 140-196

17 Pocahontas No. 3 (50) 0.10-1.40 125-187

18 Pocahontas No. 2 (10) 0.03-1.49 29-184

Due to the large size of the full field scale models (element-wise) and the added complexity

of the multiple seams structure, in order to reduce overall simulation time two different groups of

models were set up for the purposes of this investigation. A larger model (L) that includes all 20

wells within the study area and assumes a 16-layer geologic structure where seams 9, 10 and

Page 100: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

86

seams 16, 17 were aggregated into single zones; a smaller model (S) incorporates the three

selected injectors and the offset well, DD8A, and assumes an 18-layer structure. The input

parameters and material properties shown in Table 12 were assembled utilizing data from coal

sample analysis, the literature and through personal communication with Pashin (2013).

Table 12 - Input Parameters for the Combined and Four-Well Models

Number of Grid Blocks per Direction (combined models, (L)) I= 98, J= 69, K=16

Number of Grid Blocks per Direction (four-well models, (S)) I= 75, J= 39, K=18

Grid Block Sizes per Direction (combined models, (L)), ft DI= 100, DJ= 100

Grid Block Sizes per Direction (four-well models, (S)), ft DI=50, DJ=50

Coal (Matrix) Porosity (combined models, (L)), % Layers: 1 to 6, 2

Layers: 7 to 16, 1

Dual porosity model

Coal (Matrix) Porosity (four-well models, (S)), % Layers: 1 to 6, 2

Layers: 7 to 18, 1

Dual porosity model

Cleat (Fracture) Porosity (combined models,

(L)), %

Layers: 1 to 6, 0.2

Layers: 7 to 16, 0.1

Dual porosity model

Cleat (Fracture) Porosity (four-well models,

(S)), %

Layers: 1 to 6, 0.2

Layers: 7 to 18, 0.1

Dual porosity model

Matrix Permeability, mD Not in use, single permeability model

Cleat Permeability (combined models, (L)), mD Layers: 1 to 10, 30,

Layers: 11 to 16, 10,

Cleat Permeability (four-well models, (S)), mD Layers: 1 to 10, 30,

Layers: 11 to 18, 10,

Fracture Spacing per Direction, ft I: 0.01667 J: 0.01667, K: 0.01667

Well Radius, ft 0.13124

Formation Compressibility

(Matrix & Fracture), 1/psi

2E-08 (at reference pressure 416.15 psi)

(Harpalani, 2012)

Reservoir Temperature, oF 69.8

Water Density, lb/ft3 62.4779 (at reference pressure 416.15

psi)

Relative Permeability Curves

Gas-Water

Figure 43

Adsorption Isotherms according to Langmuir.

Langmuir Constants

VL, scf/ton; PL, psi

Methane: All Layers

VL= 720.003, PL= 250

Carbon Dioxide: All Layers

VL= 830, PL= 300

(VCCER, 2011)

Coal Density, lb/ft3 82.65 (VCCER, 2011)

Methane Desorption Time, days 20

Pore Pressure Gradient, psi/ft 0.315

Matrix Pressure, psi Variable, 173-717

Page 101: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

87

Coal (matrix) porosity and cleat (fracture) porosity are key input parameters determining the

initial volumetrics of the models and consequently affecting the ultimate methane recovery of the

field. High values of matrix porosity allow for less coal volume for gas adsorption on the coal

surface and thus translate to lower values of initial gas in place. High fracture porosity values

suggest higher initial water in the system and higher pressure. Due to a lack of specific

information on reservoirs conditions regarding both the matrix and the fracture porosity, porosity

values were selected for the reservoir models in order to match the volume of the historic gas

production of the field to date and to be in accordance with the estimated ultimate recovery

predicted through decline curve analysis (VCCER, 2013). Cleat (fracture) porosity was assumed

to be 10 percent of the coal (matrix) porosity. After personal communication with Pashin (2013),

it was suggested that there is a linear decrease of matrix and fracture porosity values with depth.

For this reason a coal matrix porosity of 2 percent was assumed for the first 6 shallower seams in

both the combined and four wells models and 1 percent for the deeper seams.

Langmuir type isotherms were employed in the models to represent methane and carbon

dioxide adsorption isotherms for coalbeds. For the specific injection site in Buchanan County,

there are currently no data available on the adsorptive properties of coals. For the Russell County

injection site, which is seven miles to the south-east of the current site, coal samples were sent to

Pine Crest Technologies and the methane and carbon dioxide adsorption isotherms for three coal

seams, Pocahontas No.3, Pocahontas No.7 and Pocahontas No.11, were determined (VCCER,

2011). Due to the close proximity of the two sites, an assumption was made that the coals would

have similar properties, and by taking into account that the coals in Buchanan County have

slightly higher gas contents, uniform per layer adsorption isotherms for methane and carbon

dioxide were assigned to the models.

Pore pressure is a critical parameter in modeling CBM production because it essentially

drives production. An initial pore pressure gradient of 0.315 psi/ft was used based on historical

field data (VCCER, 2013) to reflect under-saturated conditions in the study area. The

corresponding pressure values vary between 173-717 psi, depending on the depth of the coal

seam in the reservoir.

As already noted, a single cleat permeability model is employed during the simulation of

coalbed reservoirs (CMG, 2003). The stress field in the study area suggests similar fracture

permeability values along the face and butt cleats (Pashin, 2013), and therefore isotropic cleat

Page 102: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

88

permeability was assumed in the models. Cleat permeability values were determined through

history matching the gas production of the field since no other laboratory or field data for the

Buchanan injection site were available. It is believed that cleat permeability declines

exponentially with depth; as the depths in the study area vary between approximately 1000 and

2000 feet it is assumed that cleat permeability varies between 70 mD and 10 mD (McKee et al.,

1988). In the models, cleat permeability towards the lower end of the suggested range for

moderate analysis was selected; 30 mD was assigned to the first 10 seams and it was reduced to

10 mD for the 8 deepest seams.

Gas and water relative permeability curves were unavailable. Relative permeability was

assumed to be similar to that used by Mavor and Robinson (1993) when evaluating pressure

transient data for the San Juan coals based on initially developed gas and water relative

permeability curves by Gash for Fruitland coals (Gash et al., 1993). The relative permeability

curves were adjusted for the specific study area through history matching gas and water well data

(Figure 43).

Operational constraints are also an important modeling parameter. Initially the models were

set to primarily match the historic gas rate up to year 2013, while having a secondary constraint

of a minimum bottom-hole pressure of 28 psi. Even though dewatering of the coal seams is the

driving force for primary methane production from coalbed wells, as previously mentioned the

quality of the reported water data per well is unknown and, therefore, water production was

taken into account as a secondary check point for the history matching exercise. It should be also

noted that historical production data for gas and water were aggregated on a monthly basis.

Page 103: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

89

Figure 43 - Relative Permeability Curves to Water and Gas as Determined Through

History Matching Gas Production for All the Wells in the Study Area

Results and Discussion

Two different types of models are investigated in this paper. A larger model (L) representing

the whole study area incorporating 20 wells and a smaller four-well model (S) focusing on the

area around the three selected injection wells and the closest offset well, DD8A.

The main objective of the reservoir simulation efforts is to model all phases throughout the

estimated life of the wells within the study area and assess the storage capacity of coalbeds and

the extent of CO2 plumes. More specifically, the first step is to model the initial period up to year

2013, utilizing available historical gas and water production data for all wells in order to

calibrate the models. The next step is to model the one-year injection period during which it is

planned to inject 20,000 tons of carbon dioxide into the three selected wells. Finally, the model

will be used to forecast the behavior of all wells in the study area post injection until the end of

their estimated life, year 2052.

Page 104: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

90

Figure 44 - Gas Rate at Surface Conditions for the Three Injectors, DD7, DD7A, and DD8

throughout History Matching, Injection and Forecasting to Year 2023

S Models

The small models (S) incorporated the three selected injectors: DD7, DD7A and DD8 and

the offset well DD8A. The drainage area of the 18-seam geologic structure for the (S) models

was 167.9 acres and, based on the input parameters, the original gas in place (OGIP) was

estimated to be 1.84e6 standard Mcf and the original water in place 2.64e4 standard bbl.

Three different scenarios with regard to well characteristics were examined (S1, S2, S3). In

scenario S1, all 18 coal seams were perforated but not hydraulically stimulated; in scenario S2 a

negative skin was assigned to all four-wells to account for enhanced flow around the wellbore;

and in scenario S3, the hydraulic fractures were explicitly modeled.

It is common practice in reservoir modeling to assign a negative skin factor to a well in

order to account for its hydraulic fracture stimulation and, at the same time, avoid explicit

modeling of the fractures, which is very computationally intensive. Therefore, to calculate the

skin factor in this modeling work, equation (29) was initially used (Karacan, 2013).

HM INJ FOR

Page 105: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

91

(

) (29)

where

α: Function of dimensionless fracture conductivity (FCD)

xf: Fracture half-length (ft)

rw: Well radius (ft)

“α” in the equation is a function of dimensionless fracture conductivity (FCD). NSI

Technologies (2001) and Cunningham et al. (2003) reported that optimum well productivity

occurs when FCD is around 2, and is calculated by using the expression below:

(

)

(30)

where

kf: Fracture permeability (mD)

wf: Fracture width (m)

k: Permeability (mD)

xf: Fracture half-length (m)

“α” was determined as 0.3 from the dimensionless fracture conductivity versus alpha plot

(Meyer, 2012) by assuming FCD as 2. Skin factor was calculated as -6.68 by assuming a 350 feet

half-fracture length. However, when a skin factor of -6.68 is used in the GEM/CMG simulator, a

negative well productivity index is calculated. The well productivity index for a phase should be

positive and it is a function of fracture permeability, well drainage effective radius, well radius

and skin factor (GEM, 2013). For the specific reservoir properties assumed in the modeling, the

maximum skin factor that can be assigned to sustain a positive productivity index is -4.

Regarding hydraulic fractures, it was suggested by the operator of the wells that only 40

percent of the coal seams per completion stage have been successfully hydraulically fractured

(CNX, 2013) and that the thickest seams are expected to have received most of the fractures

(Pashin, 2013). For this reason, model S3, where hydraulic fractures are explicitly simulated,

only includes layers 4, 6, 8, 12, 14, 16 and 18, which represent the thickest seams and account

for 38 percent of the stimulated seams. A primary fracture width of 0.01042 feet (EPA, 2004)

with a primary fracture permeability of 10,000 mD and 350 feet of half-length is assigned in the

models. These input parameters for the hydraulic fractures translate into pseudo-fractures with

Page 106: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

92

proper grid refinement that account for non-Darcy flow with an effective permeability of 82.1

mD. Hydraulic fractures were developed along the maximum horizontal stress direction in the

area and thus the modeling grid was rotated to orient the maximum horizontal stress in the I

direction. In addition, it was assumed that the permeability in the I direction is equal to butt cleat

permeability and that in the J direction equal to face cleat permeability (Vasilikou et al., 2013).

Initially, for all three investigated scenarios S1, S2, S3 (base, skin, hydraulic fractures),

history matching of the gas rate was achieved up to year 2013. Cumulative gas production of the

field was similar for all scenarios at approximately 719.65 standard MMcf, accounting for almost

40 percent primary recovery of the original gas in place. The main coal seams contributing to gas

commingle production in all three scenarios are listed below in order (from the largest

contribution to the least): seams number 9, 15, 4, 16, 13, 8, 11, 7, and 14.

The second phase of the modeling was the injection period. The injection start date was set

for the 1st of May 2013 with injection on-going until the 1st of May 2014. Wells DD7, DD7A

and DD8 were shut in for one-month prior to commence of the injection and then turned into

injectors. The same amount of carbon dioxide was injected into DD7, DD7A and DD8,

approximately 6,667 tons per well, at a constant daily rate of 18.26 tons/day over a 1-year period.

After injection of 20,000 tons of CO2 was completed the three injectors remained shut in for one

year until the 1st of May of 2015. Gas production for all four wells in the study area was

projected to year 2023.

From the simulation results it is shown that for the three different scenarios where CO2 is

primarily injected into the coal seams there is a higher differential between the injection pressure

at the wellhead and the flowing pressure yielded at the end of the history matching period at the

specific coal seam depth. As shown in Figure 45 for scenario S3, more CO2 is injected in coal

layer 9 and that was also the seam most depleted during the history matching exercise. The

second most depleted seam during history matching was layer 15. However, during injection

layer 4, which is at a shallower depth and thus has a lower pore pressure, the seam takes up more

of the injected CO2 (Figure 45).

Page 107: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

93

Figure 45 - (A) Cumulative Gas Production per Layer Up to Year 2013 and (B) CO2

Adsorption per Layer during Injection and Post Injection Up to Year 2023 for Scenario S3

Scenario S3 (where hydraulic fractures were explicitly simulated) shows that the CO2

plumes are more extensively developed compared to scenarios S1 and S2 (skin and base

models), as shown in Figure 46, for the stimulated layers. However, since only 40 percent of the

(a)

(b)

Page 108: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

94

coal seams are hydraulically stimulated the highest levels of CO2 breakthrough at the offset well,

DD8A, are exhibited in scenario S2 (skin scenario), where an enhanced flow zone is assigned

along the wellbore for all the coal seams in the model.

Figure 46 - CO2 Adsorption Profile in gmole/ft

3 at Layer 4 for Scenarios (A) S1 Base, (B) S2

Skin and (C) S3 Hydraulic Fractures Scenario

As shown in Table 13, the results of the S model simulations show that the injection wells

DD7 and DD7A produce more CO2 during flowback when brought back on-line a year after

injection is completed in scenario S2 compared to scenarios S1 (base model) and S3 (hydraulic

fractures model). The third injector, DD8, has a higher CO2 flowback in scenario S3 (hydraulic

fractures model), Table 13. In all three scenarios, injector DD7 has a higher CO2 flowback

amount compared to DD7A and DD8; DD8 produces the least amount of CO2.

(a) (b)

(c)

Page 109: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

95

According to the modeling results of the (S) models, only 11, 6, and 10 percent of the total

20,000 tons of CO2 injected in the study area is successfully (permanently) stored up to year

2023 for scenarios S1, S2 and S3, respectively. The skin model, scenario S2, which has been

assigned the highest flow enhancement around the wellbore, is the one allowing for the least CO2

storage and maximum flowback.

As far as ultimate field methane recovery is concerned for the (S) models, 89, 94 and 90

percent (for scenarios S1, S2 and S3, respectively) of the initial gas in place is achieved. These

results are higher but still comparable to the assumption that up to more than 75 percent recovery

of methane in place can be achieved with CO2 injection in coalbeds (VCCER, 2011).

Table 13 - CO2 Flowback at the Injectors and CO2 Breakthrough at the Offset Well for the

(S) Models for Projection Time to Year 2023

CO2 Flowback at Injection Wells (Tons)

Wells Base Model (S1) Skin Model (S2) Hydraulically

Fractured Model (S3)

DD7 6,349 6,565 6,384

DD7A 5,381 5,791 5,526

DD8 5,093 4,929 5,175

CO2 Breakthrough at Offset Well (Tons)

DD8A 963 1,462 905

CO2 Total Field

Production

(Tons)

17,786 18,747 17,990

Table 14 - Cumulative CH4 Production for the (S) Models for the Base (S1), Skin (S2) and

Hydraulic Fractures (S3) Scenarios for Projection Time to Year 2023

Cumulative CH4 Production (Std Mcf)

Wells Base Model

(S1)

Skin Model

(S2)

HF Model

(S3)

DD7 578,811 639,177 584,588

DD7A 231,607 262,101 237,485

DD8 395,162 422,485 401,668

DD8A 206,529 257,726 204,616

Page 110: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

96

L Models

The larger models (L) incorporated all wells within the study area. The drainage area of the

16-seam geologic structure for the (L) models was 1552 acres. Based on the input parameters,

the original gas in place (OGIP) was estimated to be 1.77e7 standard Mcf and the original water

in place to be 2.50e5 standard bbl.

The L models were first calibrated to match gas and water production data up to 2013. Then

two different injection scenarios were applied in the simulations in order to achieve injection of

20,000 tons of CO2 in wells DD7, DD7A and DD8. More specifically, in the first injection

scenario (L1), 20,000 tons of CO2 were equally distributed among the three injectors

(approximately 6,667 tons per well) and CO2 was injected at a constant daily rate of 18.26

tons/day for 365 days, in the same manner as in the (S) models. In the second injection scenario

(L2), the same amount of CO2 was injected per well as in scenario L1, but at a higher daily rate

of 27.78 tons/day for the first 20 days of each month for a year; the injection wells were shut in

during the final days of each month during injection. Figure 47 shows the CO2 mass injection

rate at surface conditions for injection scenarios L1 and L2 over time.

After the one-year injection period was completed, different cases were examined where the

shut in period of the three injectors varied in time. The performance of all the wells within the

study area was projected to year 2052.

Page 111: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

97

Figure 47 - Gas Mass Rate of Injected CO2 at Surface Conditions Versus Time for the

Injection Scenarios L1 and L2

Four cases with varying injection scenarios (L1, L2) and shut in periods post injection for

wells DD7, DD7A and DD8 were examined. It must be noted that in all cases the well

characteristics were kept constant and no skin factor or explicit simulation of hydraulic fractures

was employed in the (L) models. In case L1a, injection scenario L1 was applied and the three

injectors were shut in for 1 year post injection; in case L1b, injection scenario L1 was used and

the injectors were shut in for 4 years; in case L2a, injection scenario L2 was employed and the

selected injection wells were shut in for 1 year post injection; and in case L1c injection scenario

L1 was applied and the injectors were shut in for 36 years post injection, which means they were

shut in throughout the projected time of the simulation.

During injection, for both injection scenarios L1 and L2, the bottom hole pressures in the

wells did not exceed the maximum allowable pressure as indicated by a U.S. EPA class II

UIC permit (VCCER, 2013). For injection scenario L1, the maximum bottom hole

pressures for the three injectors DD7, DD7A and DD8 were approximately 423, 484 and

416 psi respectively; for injection scenario L2, the maximum bottom hole pressures were

482, 564 and 483 psi respectively. Also, under the specific pressure and temperature

Page 112: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

98

conditions, carbon dioxide was injected as a gas and remained in the gaseous state

throughout the injection,

Figure 48.

Figure 48 - Pressure-Temperature Phase Diagram for CO2 and Bottom Hole Pressures at

Injection Well DD7 for Injection Plans L1a, L1b, and L1

Page 113: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

99

Figure 49 - Cumulative Gas Production of All the Wells in the Study Area for All Cases

Examined for the (L) Models

The main objective of the modeling work as mentioned previously is to assess CO2 storage

capacity of coalbeds, examine the extent of CO2 plumes and investigate the potential of

enhanced gas recovery at the injectors and offset wells.

The cumulative gas production per well in the study area for all different cases (L1a, L1b,

L2a, L1c) is presented in Figure 49. To estimate enhanced gas recovery of the wells, their

cumulative gas production determined by the simulations is compared to the estimated ultimate

recovery based on decline curve analysis. In more detail, Cardno MM&A (2013) forecasted the

ultimate primary gas production of the wells within the study area without taking into account

CO2 injection, based solely on exponential decline curve fitting of historic gas production data

(VCCER, 2013). As shown in Figure 49, for cases L1a, L1b and L2a cumulative gas production

of wells CC6, CC6A, CC7, DD5, DD6, DD6A, DD8A, EE6 and EE7 is higher compared to the

estimated gas production via decline curve analysis by approximately 2, 22, 53, 33, 17, 35, 15,

49 and 10 percent, respectively. For case L1c, where all three injectors remain shut in post

injection until the end of the forecasted period, the cumulative gas production at wells CC7,

DD5, DD6, DD6A, DD8A, EE6 and EE7 is lower than in cases L1a, L1b and L2a but it is still

Page 114: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

100

enhanced in comparison to the predicted ultimate recovery of these wells without CO2 injection.

It must be noted that in all cases, L1a, L1b, L2a, L1c, cumulative gas production for wells CC8,

CC9B and EE9 is significantly higher than what is predicted through decline curve analysis. This

can be partially explained because the average pore pressure for these wells is higher due to a

higher depth to the gas producing seams. In addition, the proximity of these wells to the no-flow

boundary of the model should be further investigated to determine whether it contributes to the

higher gas production.

Regarding CO2 storage in the coal seams, it has been determined by the modeling work for

cases L1a, L1b, L2a, L1c, as shown in Table 16, that 8.15, 18.2, 8.6, 81.64 percent of the total

20,000 tons were successfully sequestered up to year 2052. Comparing the two injection

scenarios that have the same shut in periods for the three injectors post injection, it results that

performing a type of “huff and puff” injection (scenario L2) yields slightly higher storage of

CO2. With respect to case L1b, where the continuous injection plan L1 was applied and the three

injectors were shut in for 4 years, less CO2 as compared to cases L1a and L2a has flowed back

and broken through in total. More specifically, approximately 16,354 tons of CO2 versus 18,371

and 18,359 tons of CO2 were produced in cases L1a, L1b and L2a, respectively. Out of all the

cases investigated, the most successful regarding CO2 storage - but the one with the least

enhanced gas recovery - is case L1c, where the three injectors remain shut-in throughout the

projected period up to year 2052.

Finally, results of the (L) models showed that ultimate field methane recovery is about 73

percent of the initial gas in place for cases L1a, L1b, L2a. For case L1c it was approximately 67

percent of the original gas in place, six percent lower than what the other models predicted.

Page 115: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

101

Table 15 - Cumulative Gas Production (CGP) for All Wells for Cases, L1a, L1b, L2a, L1c

Different Injection Plans & Injectors Shut in Periods Models Baseline

Wells

CGP

(Std Mcf)

L1a

CGP

(Std Mcf)

L1b

CGP

(Std Mcf)

L2a

CGP

(Std Mcf)

L1c

EUR

(Std Mcf)

CC6 987,936 994,560 987,979 954,294 971,235

CC6A 413,518 421,162 413,518 338,075 340,920

CC7 732,541 744,980 732,609 722,581 481,097

CC7A 276,046 283,501 276,089 277,727 339,373

CC8 1,422,654 1,439,308 1,422,845 1,366,867 607,132

CC9 36,651 36,651 36,651 36,651 N/A

CC9A 709,577 715,687 709,671 659,101 N/A

CC9B 629,022 633,463 629,079 580,365 320,446

DD5 720,020 723,772 720,039 682,670 588,601

DD6 1,135,585 1,146,610 1,135,673 1,102,503 972,886

DD6A 341,131 348,545 341,132 327,608 252,699

DD7 534,744 491,058 534,361 371,580 638,375

DD7A 242,935 201,746 242,843 90,420 277,865

DD8 419,415 362,923 418,783 207,072 370,263

DD8A 327,622 329,922 328,006 306,586 285,564

DD9 904,661 914,451 904,687 877,530 1,105,313

DD9A 447,732 452,743 447,714 421,883 512,732

EE6 737,057 748,151 737,105 718,395 530,887

EE7 914,234 925,790 914,227 884,206 828,465

EE8 666,874 683,045 666,752 670,343 713,174

EE9 360,081 365,278 360,108 335,567 161,065

Page 116: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

102

Table 16 - CO2 Flowback and Breakthrough for Cases, L1a, L1b, L2a, L1c

Different Injection Plans & Shut in Periods Models

Wells Produced CO2

(Tons)

L1a

Produced CO2

(Tons)

L1b

Produced CO2

(Tons)

L2a

Produced CO2

(Tons)

L1c

CC6 4.95E-04 2.50E-02 9.19E-04 1.89E-01

CC6A 4.90E+00 2.23E+01 4.92E+00 6.58E+01

CC7 4.22E+00 1.72E+01 4.50E+00 5.79E+01

CC7A 5.52E-02 2.14E-01 5.88E-02 6.24E-01

CC8 1.08E-02 1.55E-01 1.61E-02 7.56E-01

CC9 0.00E+00 0.00E+00 0.00E+00 0.00E+00

CC9A 2.30E-05 2.08E-05 2.59E-05 1.87E-05

CC9B 2.70E-05 2.70E-05 2.78E-05 2.69E-05

DD5 4.16E-06 2.21E-06 3.25E-06 1.47E-06

DD6 3.10E+01 1.13E+02 3.16E+01 2.50E+02

DD6A 8.79E+00 2.54E+01 8.89E+00 5.44E+01

DD7 5.92E+03 4.70E+03 5.91E+03 0.00E+00

DD7A 5.78E+03 4.93E+03 5.77E+03 0.00E+00

DD8 5.54E+03 4.64E+03 5.56E+03 0.00E+00

DD8A 1.03E+03 1.74E+03 1.01E+03 2.85E+03

DD9 1.12E-01 7.13E-01 1.20E-01 2.21E+00

DD9A 7.60E-05 3.62E-04 7.68E-05 1.44E-03

EE6 2.25E-02 1.54E-01 2.32E-02 9.32E-01

EE7 2.86E-03 9.71E-03 2.84E-03 2.16E-03

EE8 6.44E+01 1.61E+02 6.41E+01 3.86E+02

EE9 9.99E-06 1.30E-05 1.20E-05 1.00E-05

Page 117: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

103

Conclusions

In this paper the modeling procedure for coal bed methane production and carbon dioxide

injection from and into multiple seams through selected wells in Buchanan County, VA, is

discussed in detail. Initially, well performance is modeled using three different modeling

approaches. Results show that when a negative skin factor is assigned to the production wells,

primary production during history matching, flowback of CO2 at the injectors and post injection

breakthrough at offset wells is possibly overestimated. When explicitly simulating hydraulic

fractures, modeling is more computationally intensive but results appear more moderate

compared to the case of the negative skin factor. Furthermore, when modeling the behavior of

hydraulically fractured seams, the extent of the CO2 plumes is more representative of actual field

conditions. Nevertheless, explicit representation of hydraulic fractures introduces further

modeling uncertainties on fracture parameters such as the actual width, length, effective

permeability and flow properties.

The process of CO2 injection into coal seams with the objective of assessing the potential

enhanced gas recovery and the permanent storage of CO2 was also examined.

Two different injection scenarios were examined. In the first scenario CO2 was injected at a

constant rate throughout the entire injection period and in the second scenario a small huff and

puff injection test was performed instead, but at a higher injection rate. Results of the second

scenario show that permanent CO2 storage is better.

Furthermore, it was concluded CO2 permanent storage is proportional to the time interval for

which the injection wells are shut in post injection. Maximum CO2 storage occurred when the

wells were not returned to production. As a result, however, enhanced gas recovery of the field is

reduced for wells with higher post injection shut in times.

Acknowledgements

Financial assistance for this work was provided by the U.S. Department of Energy through

the National Energy Technology Laboratory’s Program under Contract No. DE-FC26-

04NT42590 and DE-FE0006827.

Page 118: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

104

References

CNX Gas Company LLC, 2013. Personal communication.

Computer Modeling Group Ltd (CMG), 2003. GEM. Calgary, Canada.

Conrad, J. M., Miller, M. J., Phillips, J., Ripepi, N. 2006. Characterization of Central

Appalachian Basin CBM Development: Potential for Carbon Sequestration and Enhanced

CBM Recovery, 2006 International Coalbed Methane Symposium, Preprint 0625,

Tuscaloosa, AL.

Cunningham, L., Smith, M., Miller, B., Britt, L., and Vandersypen, R. 2003. Geoscientists

Provide Critical Needs for Optimal Frac Design. Houston Geological Society Bulletin 45:5,

45-46.

US EPA. 2004. Evaluation of Impacts to Underground Sources June 2004 of Drinking Water by

Hydraulic Fracturing of Coalbed Methane Reservoirs (Final Report). U.S. Environmental

Protection Agency, EPA/816/R-04/003.

Gash, B.W., Volz, R.F., Potter, G., and Corgan, J. M. 1993. The effects of cleat orientation and

confining pressure on cleat porosity, permeability and relative permeability in coal. Paper

9321 in Proceedings of the 1993 International CoalBed Methane Symposium. Tuscaloosa:

University of Alabama.

Grimm, R. P. 2010. Seal Evaluation of the Hensley Shale for Carbon Dioxide (CO2) Enhanced

Coalbed Methane Recovery in the Breathitt Group, Nora Field, Virginia. Geological Society

of America Abstracts with Programs, 42, 137.

Harpalani, S. 2012. Personal communication.

Karacan, C.O. 2013. Production History Matching to Determine Reservoir Properties of

Important Coal Groups in the Upper Pottsville formations, Brookwood and Oak Grove

fields, Black Warrior Basin, Alabama. Journal of Natural Gas Science and Engineering, 10,

51-67.

Karmis, M., Ripepi, N., Miskovic, I., Conrad, M., Miller, M., and Shea, C. 2008. CO2

Sequestration in Unminable Coal Seams: Characterization, Modeling, Assessment and

Testing of Sinks in Central Appalachia. Twenty-Fifth Annual International Pittsburgh Coal

Conference, Pittsburgh, PA, USA, 2008. CD-ROM of Proceedings, Paper # 29-3.

Page 119: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

105

Liu, S., Harpalani, S., and Pillalamarry, M. 2012. Laboratory measurement and modeling of coal

permeability with continued methane production: Part2 - Modeling Results. Fuel, 94, 117-

124.

Mavor, M.J.and Robinson, J.R. 1993. “Analysis of Coal Gas Reservoir Interference and Cavity

Well”, Paper SPE, 25860, presented at the Joint Rocky Mountain Regional and Low

Permeability Reservoirs Symposium, Denver, CO, April 26-28.

McCulloch, C.M., Deul, M., and Jeran, P.W. 1974. Cleats in bituminous coalbeds. U.S. Bur.

Mines, Rept. Invest. 7910, 23.

Mckee, C.R., Bumb, A.C., and Koenig, R.A. 1988. Stress-Dependent Permeability and Porosity

of Coal. SPE-12858-PA, 3, 81-91.

Meyer Fracturing Simulators. 2012. User’s Guide Tenth Edition.

NSI Technologies Inc. 2001. Optimum Fracture Design. Fract Tips. 3:1, 1-2.

Pashin, J. 2013. Personal communication.

Ripepi, N., Karmis, M., Miskovic, I., Shea C., and Conrad, J.M. 2009. Results from the Central

Appalachian Basin Field Verification Test in Coal Seams. 26th Annual International

Pittsburgh Coal Conference, Pittsburgh, PA, USA, 2009. CD-ROM of Proceedings, Paper #

33-4.

Vasilikou, F. C., Keles, Agioutantis, Z., Ripepi, N. and Karmis, M. 2013. Model Verification of

Carbon Dioxide Sequestration in Unminable Coal Seams with Enhanced Coal Bed Methane

Recovery, 23rd World Mining Congress, August 11-15, Montreal, Canada.

Virginia Center for Coal and Energy Research (VCCER). 2011. Final Technical Report:

Characterization and Field Validation of the Carbon Sequestration Potential of Coal Seams

in the Central Appalachian Basin, 2011: Blacksburg, VA.

Virginia Center for Coal and Energy Research (VCCER). 2013. 6th Quarterly Report: Injecting

Carbon Dioxide into Unconventional Storage Reservoir in the Central Appalachian Basin,

with an Emphasis on Enhanced Coalbed Methane Recovery, to Validate Prior Geological

Characterization, 2013: Blacksburg, VA.

Virginia Center for Coal and Energy Research (VCCER). 2014. 7th Quarterly Report: Injecting

Carbon Dioxide into Unconventional Storage Reservoirs in the Central Appalachian Basin,

with an Emphasis on Enhanced Coalbed Methane Recovery, to Validate Prior Geological

Characterization, 2014: Blacksburg, VA.

Page 120: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

106

Virginia Department of Mines, Minerals and Energy (DMME). “Well Records: Completion

Report”. Division of Gas and Oil Data Information System. DGO E-Forms Center. Web. 21

Dec 2011.

Virginia Department of Mines, Minerals and Energy (DMME). “Monthly Gas Production by

Well”. Division of Gas and Oil Data Information System. DGO E-Forms Center. Web. 21

Dec 2013.

Page 121: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

107

SUMMARY AND CONCLUSIONS

Sequestration of carbon dioxide (CO2) into unmineable coal seams has been underway for

several years as a way to mitigate the greenhouse effect with a potential of economic prosperity

related to enhanced gas recovery. The Virginia Center for Coal and Energy Research (VCCER),

part of the Coal Seam Group of the Southeast Regional Carbon Sequestration Partnership

(SECARB), successfully completed a small scale injection project in 2009 in Russell County,

VA, where 1000 tons of CO2 were injected into multiple thin coal seams through one vertical

well over a period of one month. In 2014, a larger scale sequestration project is scheduled, where

20,000 tons of CO2 will be injected into three vertical coalbed methane wells in a coal field in

Buchanan County, VA, over a one year period. The main objectives for these injection tests are

to assess storage capacity of “stacked” coal seams, enhance understanding of the physical and

mechanical processes taking place and examine the potential of enhanced gas recovery.

During primary coalbed methane production and enhanced production through CO2

injection, a series of complex physical and mechanical phenomena occur. The ability to represent

the behavior of a coalbed reservoir as accurately as possible via computer simulations yields

insight into the processes taking place and is an indispensable tool for the decision process of

future operations. The economic viability of projects can be assessed by predicting production,

well performance can be maximized, drilling patterns can be optimized and, most importantly,

associated risks with operations can be accounted for and potentially avoided.

Simulations require a large number of input parameters and high computational capabilities

in order to accurately predict the behavior of the reservoir. Therefore, in current modeling

practices many simplifying assumptions need to be employed. The shortcomings of the modeling

approaches specific to coalbed reservoirs are:

1. In areas with complex geologic structures, it is common practice to use a simplified

approach of aggregating the coal seams into zones to reduce the complexity of the

model and thus the required computational intensity. However, important

information with respect to the commingled production and the injection mechanism

is not accounted for in this method.

Page 122: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

108

2. Often to reduce modeling time in the simulation of large multi-well areas, separate

single-well models are investigated, disregarding well interference of the field - one

of the key parameters in coalbeds.

3. In order to minimize the computational intensity, well stimulation is either not

accounted for or not representative of the in situ conditions for complex reservoirs.

4. There is an infinite solution space for history matching exercises due to the large

number of unknown vs. known input parameters. Often, the calibration of the models

is not constrained within an acceptable range.

In this dissertation the aforementioned shortcomings were addressed and the initial reservoir

model for the Russell County site was updated. Subsequently the reservoir model for the

Buchanan County test site was constructed and the preliminary simulations for CO2 sequestration

and enhanced gas recovery were conducted. The following were accomplished in this work:

1. Sensitivity analysis was conducted for a number of model input parameters and the

key parameters and their effect identified.

2. The dynamic evolution of permeability during primary and enhanced recovery from

coalbeds, which is extensively referred to in the literature, was investigated. Coupled

flow and geomechanical simulations were developed to assess the significance of

implementing permeability changes into full field scale simulations.

3. Well stimulation approaches, including a negative skin factor and explicit simulation

of hydraulic fractures, were considered and compared.

4. Different CO2 injection scenarios into multiple seams for a multi-well field area were

modeled and the potential of enhanced gas recovery was assessed.

Analytical and numerical models proposed in the literature to address the phenomenon of

dynamic evolution of permeability during primary and enhanced recovery in coalbeds were

critically assessed. These models have been only used for single seams, single well and small

area representations. An algorithm was developed to couple a reservoir simulator with a

geomechanical code to examine permeability changes during methane production and CO2

injection for a single coal seam. It was concluded that for relatively flat seams, where there are

no areas of large stresses and strains localization, permeability changes are not significant and

thus it was decided to not be considered in the full-field scale simulations.

Page 123: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

109

Different modeling approaches to account for well stimulation were implemented and

compared. From the analyses it was shown that assigning a negative skin factor, which is

determined through history matching, often overestimates the enhanced flow properties around

the wellbore and facilitates post injection flowback of CO2 at the injector. Explicit simulation of

hydraulic fractures can be controlled to be more representative of in situ field conditions,

although more unknowns are introduced in the simulation: primary width, primary permeability,

effective half-length, fracture orientation and flow regime properties. Not accounting for well

stimulation can lead to unrealistic modeling results. For example, such models could not predict

that during injection in the Russell County test site there would be CO2 breakthrough at the

monitoring well closest to the injector within hours of starting injection. More realistic modeling

results were obtained where hydraulic stimulation of the injection well was included in the

simulation model.

Two CO2 injection scenarios for the Buchanan County, VA, site were examined. In the first

scenario approximately 6,667 tons of CO2 were injected into 18 coal seams at each injector at a

constant rate for a year; in the second scenario a “huff and puff” type of CO2 injection at a higher

rate for the first twenty days per month for a year was applied. It was concluded that for the first

scenario there was slightly higher CO2 breakthrough at the injectors compared to the second

scenario. In the second scenario there were intermediate time intervals during the one-year

injection, allowing CO2 to set in. From the analysis it was also shown that the time the injectors

are shut in post injection is critical to the percentage of CO2 successfully stored. The longer the

wells are shut in the less CO2 flows back at the injectors. Maximum CO2 storage can be achieved

when the injector well are not returned to production. CO2 breakthrough at offset wells for all

injection scenarios was also noted, but it was significantly less compared to flowback.

Enhanced gas recovery at the injectors and offset wells was noted in the modeling work. It

was concluded that since the majority of the injected CO2 flowed back, the primary mechanism

for enhanced recovery for the coals in the study area is not due to the CO2 preferential adsorption

by the coal matrix and CH4 displacement, but is because of the “renewal” of the pressures in the

reservoir.

The most important conclusion for the reservoir simulation work is that it is not a

straightforward process. Even though the models can provide reasonable solutions based on the

Page 124: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

110

available data, it is imperative that current models are updated with new data from laboratory

testing and/or field work monitoring to ensure that they remain stable and robust.

Detailed results of the aforementioned are presented in the papers included in the main body

of this work.

Further Work

As discussed, the main problem in reservoir modeling is the uncertainty in the magnitude

and range of several input parameters. Following extensive laboratory testing related to coalbed

parameters over a number of years, reservoir model input parameters have been established;

however, there are still gaps in the understanding of physical processes. For improved

predictions through reservoir simulations, field parameters need to be determined more

accurately and at different times throughout the lifetime of a reservoir.

For instance, it is necessary to achieve better monitoring of the quantity of the water that is

produced from each well during production. This would result in better history matching of the

gas produced. In addition, production pressures at the well head and/or the bottom of the well

should be accurately recorded.

Also, in order to better understand and “decode” commingled production from multiple

seams there is a need to perform field tests where a seam or a group of seams are isolated and

properties such as permeability can be determined via transient pressure testing.

Camera logging of the injectors is proposed to obtain more information regarding hydraulic

fracture simulation of the well and create more representative models. In this way the modeling

efforts will better capture CO2 flow and potential breakthrough and storage processes.

There is a constant need to use reservoir simulators with enhanced capabilities so that less

simplifying assumptions are required for the models and more details are incorporated. The next

step would be to use a higher end simulator such as Eclipse by Schlumberger, where ability to

model more geologic details and create larger element models is provided.

Finally, incorporating a temperature gradient in the models to better account for potential

phase changes of CO2 during injection and storage will be an important consideration.

Page 125: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

111

Recommendations for Transferring Lessons Learned from Coalbed Methane

Modeling to Shale Gas Modeling

Research on primary production from coalbed reservoirs has been underway for several

decades. The best practices to estimate the original gas in place are established and there is

agreement within the research community on the basic methane production mechanism. Yet

there are still questions which need to be answered, as previously mentioned, to further

understand the production mechanism from multiple thin unmineable coal seams and the

interaction with CO2 injection. Currently, there is an interest in the industry to transfer the

experiences and expertise from coalbed reservoir to “unlock” the great energy potential of the

shale reservoirs.

In both coalbed and shale reservoirs a significant portion of the gas in place is stored via

adsorption on the rock matrix. This is their main - and should be considered to be their only -

similarity. There are significant differences between the geologic properties, such as initial pore

pressure, porosity and adsorption isotherms that result in different initial volumetrics and

production mechanisms. In addition, shale gas reservoirs are usually much deeper than coalbed

methane reservoirs. The decades of coalbed methane research poses the right questions to be

answered in order to identify the critical unknown parameters in hydrocarbon exploration from

shale reservoirs. However, there should not be a direct transfer of properties from coalbeds to

shale reservoirs since it is likely to lead to the wrong conclusions. Regarding primary production

from shales, the following critical questions need to be addressed:

1. What is the effective volume activated through hydraulic stimulation, depending on

the stimulation technology employed, particularly with regard to the frac-fluid

composition? Further field characterization and monitoring is required.

2. What are the different flow regimes for the (i) primary artificial channels developed

with hydraulic stimulation, (ii) the activated dendritic pattern perpendicular to the

primary fractures, and (iii) within the shale matrix? Field determination through

pressure transient testing and further laboratory-testing accounting for in situ

conditions needs to be conducted to determine permeability ranges for each case.

3. What happens to the water used for hydraulic treatment? In the case where it is

produced, how are the relative permeability curves affected? If it stays in the

Page 126: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

112

reservoir how are the pressure profile, and the consequent gas release from the shale

formation, shaped?

4. What percentage of the original estimated gas in place is recovered in relation to the

well characteristics, such as lateral length, orientation to in situ stresses, proximity to

other wells, frequency and intensity of hydraulic stages?

5. Could the use of CO2 for well stimulation, pressure management and/or hydraulic

fracturing benefit the production of both gas and condensate, specifically the ability

of CO2 to change the viscosity of immobile condensates and allow for their

production?

There are a number of issues that need to be resolved if shale reservoirs are to be used for

permanent CO2 storage in a similar manner to deep coal seams. It has been determined that there

is a larger affinity of CO2 to the coal and the shale matrix compared to CH4, especially in shale

when there is an increase of clay content. The in situ behavior of CO2 and the coal matrix is still

under investigation through a series of small and medium CO2 sequestration projects in

unmineable coal seams. The same aspects should be examined for the case of CO2 sequestration

in shale reservoirs with the investigation of CO2 properties under the different depth and pressure

conditions of shales. In addition, there is less interference between wells and different well

development patterns when exploiting shale reservoirs; the potential of enhanced gas recovery at

offset wells should also be examined and potentially could be negligible depending on well

drilling patterns and spacing

Historic production data for coalbeds are available for in excess of 20 to 30 years, a large

portion of the assumed life of a 50-year well. Type curve fitting of the historic points and

development of decline curve analysis to estimate the behavior of new coalbed methane wells is

well established. From the curve fitting it is concluded that for production of coalbeds there is an

initial period of dewatering and increase of the gas rate, then a plateau in production due to

commingle production from multi-seams and at the end a decline period which is best fitted with

an exponential segment. The historic production data available for horizontal shale wells in most

basins are limited to production of approximately seven years. For initial production, a sharp

decline is best fitted via a hyperbolic decline segment for production of free gas in the system

and for the rest of the well life an exponential decline is assumed given that desorption from the

Page 127: MODELING CO SEQUESTRATION AND ENHANCED GAS …...SEQUESTRATION AND ENHANCED GAS RECOVERY IN COMPLEX UNCONVENTIONAL RESERVOIRS Foteini Vasilikou ABSTRACT Geologic sequestration of CO

113

shale matrix will be activated. Data are still not available to support this argument. However, it is

evident that there is a different production mechanism in shale compared to coalbed.

To understand the physical and dynamic occurring processes in shales, all the pertinent

properties need to be evaluated from the nano scale to investigate the adsorption/desorption

properties, and moving toward larger scales to bridge the knowledge of the micro to the macro

processes. Extensive field characterization is required by monitoring wells and with in situ

properties testing. Most importantly, production data must be collected over a period of time to

evaluate the basic production mechanism.


Recommended